UNISYS Relational Concepts

and SQL Programming

Student Guide

Copyright® MCMXCIII Unisys Corporation. All rights reserved.

Unisys is a registered trademark of Unisys Corporation.

Unisys Institute of Technology (UIT) North American Education

Printed in U S America
EL3605 SG 3605
 Relational Concepts and SQL Programming

Contents

Course Description xv

About This Course xvi

How to Use This Document xvi
Learning Environment xvi
Trademark Information xvi

Reference Documents xvii

Agenda xviii

Module 1: Introduction

Module Objectives 1-3

Benefit Statement 1-3

Materials 1-3

Relational Database 1-4

Operations on Tables 1-5

RESTRICT Operation 1-6
PROJECT Operation 1-7
JOIN Operation 1-8

Advantages of the Relational Model 1-9

What is RDMS 1100? 1-10

Interfaces to RDMS 1-11

Interfaces to RDMS 1-12

The Structured Query Language (SQL) 1-13

Exercise 1-1 1-14

Module 2: Retrieving Data with the SELECT Statement

Module Objectives 2-3

Benefit Statement 2-3

SG 3605 iii
Relational Concepts and SQL Programming

Materials 2-3

SQL Queries 2-4

Sample Tables for Examples and Exercises 2-5

SELECT Command 2-6

Selecting Columns 2-7
DISTINCT Values 2-8
The WHERE Clause 2-9
WHERE Clause Operators 2-10
WHERE Clause Usage 2-11
The ORDER BY Clause 2-12
Arithmetic Expressions 2-13
Order of Precedence 2-14

Exercise 2-1 2-15

Module 3: Using SQL in Application Programs

Module Objectives 3-3

Benefit Statement 3-3

Materials 3-3

Using SQL Commands in 3GL Programs 3-4

Accessing RDMS 1100 Data 3-4
3GL Interfaces 3-4
UDS Environment Commands 3-4

SQL and Unisys Extensions 3-5

Summary of Selected SQL Commands 3-6

Comparing Embedded and Interpretive SQL 3-8

Source 3-9
Compilation 3-9
Collection or Link 3-9
Execution 3-9

SQL Commands in COBOL Programs 3-10

Interpretive SQL 3-10
Embedded SQL 3-12

Command Building Blocks 3-15

Constants 3-15
Naming Database Objects 3-15

iv SG 3605
 Relational Concepts and SQL Programming

RDMS 1100/2200 Data Types 3-16

Character Data Types 3-16
Numerical Data Types 3-16
RDMS and 3GL Data Type Equivalencies 3-18

RDMS 1100/2200 Terminology 3-19

Cursor 3-19
Currency 3-19
Application Group 3-19
Thread 3-19

Program Flow 3-19

BEGIN THREAD Command 3-21

DECLARE CURSOR Command 3-22

OPEN CURSOR Command 3-23

FETCH Command 3-24

Passing Information Between a Program and RDMS 3-25

Host Program Variables 3-25
Placeholder Variables in ISQL 3-25
Embedded Variables in ESQL 3-25
Parameter Passing Limitations 3-25
ISQL - Placeholder and Host Program Variables 3-26

END THREAD Command 3-27

USE DEFAULT QUALIFIER Command 3-28

Exercise 3-1 3-29

Lab 3-1 3-31

Error Handling Choices 3-33

RDMS Error Handling 3-34
GETERROR Command 3-35
Example Using GETERROR (ACOB) 3-36
ESQL Error Handling 3-37
Data and Commands for ESQL Error Handling 3-38
WHENEVER Command 3-39
ESQL Data Area Declarations 3-40
Example Using GETERROR (UCOB) 3-41

Lab 3-2 3-42

SG 3605
Relational Concepts and SQL Programming

Module 4: More WHERE Clause Operators

Module Objectives 4-3

Benefit Statement 4-3

Materials 4-3

Testing Conditions 4-4

Multiple Conditions 4-4

NULL Values 4-5

Testing NULL Values 4-6

Logical Operators 4-7

Truth Tables for Three-Valued Logic 4-8

AND 4-8
OR 4-8
NOT 4-8

BETWEEN Operator 4-9

Order of Precedence of Operators 4-10

LIKE Operator 4-11

LIKE Examples 4-12

Exercise 4-1 4-13

Lab 4-1 4-15

Module 5: The IPF SOL Interface

Module Objectives 5-3

Benefit Statement 5-3

Materials 5-3

The IPF SQL Interface to RDMS 1100 5-4

Using IPF SQL 5-5

Invoking IPF 5-5

vi SG 3605
 Relational Concepts and SQL Programming

Sample IPF SQL Session 5-7

Creating an SQL File 5-8

Sample IPF SQL File 5-9

Executing an SQL File 5-9

Saving Query Results in a File 5-10

Lab 5-1 5-11

Module 6: The MAPPER Relational Interface

Module Objectives 6-3

Benefit Statement 6-3

Materials 6-3

MAPPER Relational Interface (MRI) 6-4

Invoking RDI 6-6

RDI Menu Selections 6-7

Environment Menu Selections 6-8

Selection Example 1 - Menu Access 6-9

Selection Example 2 - Freeform Entry 6-11

Exiting MAPPER 6-13

Embedding Commands in MAPPER Runs 6-14

MAPPER Run Example Using FCH 6-15

Lab 6-1 6-16

Lab 6-2. 6-17

Module 7: Grouping Functions

Module Objectives 7-3

Benefit Statement 7-3

SG 3605 vii
Relational Concepts and SQL Programming

Materials 7-3

Grouping Functions 7-4

AS Clause 7-5

Function Examples 7-6

Grouping Functions Usage 7-7

COUNT Function 7-8

GROUP BY Clause Usage 7-9

Grouping on More Than One Column 7-10

WHERE and GROUP BY Clauses Together 7-11

HAVING Clause 7-12

Full Query Specification 7-13

Exercise 7-1 7-14

Module 8: Joining Tables

Module Objectives 8-3

Benefit Statement 8-3

Materials 8-3

Equijoins 8-4

Join Guidelines 8-6

Multi-Table Joins 8-7

Self-Joins 8-10

Self-Join Guidelines 8-11

How RDMS Joins Tables 8-13

Exercise 8-1 8-14

viii SG 3605
 Relational Concepts and SQL Programming

Module 9: More Retrieval Considerations

Module Objectives 9-3

Benefit Statement 9-3

Materials 9-3

Data Access Considerations 9-4

Indicator Variables 94
Indicator Variable Examples 9-5
Handling Tables with More Than 28 Columns (ACOB) 9-6

Reusing Cursors 9-7

Cursor Review 9-7
Describe Reusable 9-7
Host Program Variables in Cursors 9-7
Program Flow for Reusing Cursors 9-8
CLOSE Command 9-9
DROP CURSOR Command 9-10
OPEN ... USING Command 9-11
Example of Host Variables in Cursor 9-11

Lab 9-1 9-13

Sequential and Random Processing on a Cursor 9-14

Currency and Fetching 9-14
FETCH [ option ] Command 9-15
Introduction to Random Processing on a Cursor 9-16
DECLARE CURSOR Command 9-17
LOCATE Command 9-17
FETCH Command 9-18
Random Processing Example 9-19

Lab 9-2 9-20

Retrieval Efficiency Considerations 9-21

Singleton SELECT Command 9-21
Singleton SELECT Example 9-22
FETCH NEXT n Command 9-23
FETCH NEXT n Example 9-24
How RDMS Retrieves Data 9-25
Retrieval Using No Key 9-26
Retrieval Using A Secondary Key 9-27
Secondary Keys 9-28
Accessing Data Using Keys 9-29

Exercise 9-1 9-30

SG 3605 ix
Relational Concepts and SQL Programming

Module 10: Data Manipulation

Module Objectives 10-3

Benefit Statement 10-3

Materials 10-3

Data Manipulation 10-4

BEGIN THREAD Command Revisited 10-5

ROLLBACK Command 10-6

Inserting Rows into Tables 10-7

Introduction to INSERT 10-7
INSERT Command 10-8

Changing Existing Data 10-9

Introduction to UPDATE 10-9
UPDATE Positioned Command 10-10
UPDATE Searched Command 10-11
Sample COBOL Update Program 10-12

Deleting Rows from Tables 10-16

DELETE Command 10-16 —
DELETE Examples 10-17
Sample COBOL Delete Program 10-18

Cursor Requirements for Commands 10-21

Exercise 10-1 10-22

Lab 10-1 10-23

Module 11: Keys and Updating

Module Objectives 11-3

Benefit Statement 11-3

Materials 11-3

Primary Keys 11-4

Update Anomalies 11-5

x SG 3605
 Relational Concepts and SQL Programming

Foreign Keys 11-6

Introduction 11-6
Parent Table / Child Table Relationships 11-6
Foreign Key Considerations 11-7
Foreign Keys and INSERT 11-8
Foreign Keys and DELETE 11-9
Foreign Keys and UPDATE 11-10
Rules for Parent - Child Relationships 11-11

Lab 11-1 11-12

Module 12: Obtaining Table Definition Information

Module Objectives 12-3

Benefit Statement 12-3

Materials 12-3

MAPPER RDI Access 12-4

Unisys Repository Manager (UREP 1100) 12-6

Introduction 12-6
Calling the Repository Manager 12-7
UREP Example 12-8

Lab 12-1 12-10

Module 13: Views

Module Objectives 13-3

Benefit Statement 13-3

Materials 13-3

Views 13-4
Basic Definition 13-4
What is a View? 13-5
Purpose of Views 13-6

CREATE VIEW Command 13-7

View Example 13-8
Columns in Views 13-9
Data Checking 13-10
Security 13-10

SG 3605 xi
Relational Concepts and SQL Programming

Updating Views 13-11

DROP VIEW Command 13-12

Exercise 13-1 13-13

Lab 13-1 13-14

Module 14: More About Threads and UDS

Module Objectives 14-3

Benefit Statement 14-3

Materials 14-3

Threads 14-4

Threads and Steps 14-5

Thread Control Commands 14-6

Recovery and Thread Control 14-7

BEGIN THREAD Command Revisited 14-8

Application Groups 14-9
Recovery Options 14-10

COMMIT Command 14-11

Message Recovery on COMMIT/END THREAD 14-12

Message Recovery on ROLLBACK 14-13

Locking 14-14
Concurrent Access 14-14
Locking 14-14
Types of Locks 14-15

LOCK Command 14-16

Implicit RDMS Row Locks 14-17

Locking Conflicts 14-18

Deadlock 14-19

Resolving Deadlock 14-20

xii SG 3605
 Relational Concepts and SQL Programming

UNLOCK Command 14-21

Program Flow with Locks 14-22

Exercise 14-1 14-23

Module 15: Getting Physical - Tables and Storage Areas

Module Objectives 15-3

Benefit Statement 15-3

Materials 15-3

Creating a Database 15-4

Schema 15-6

Storage-Areas 15-7

Storage-Areas, Files, and Tables 15-8

Logical Tables and Physical Files 15-9

Table Versions 15-10

Fully Qualified Table-Name 15-11

USE DEFAULT 15-12

How RDMS 1100 Locates Data 15-14

How RDMS 1100 Resolves Table References 15-15

Creating Versions Example 15-16

Exercise 15-1. 15-17

Appendix A: System Usage Information

System Essentials A-3

Demand Session and Instructions A-5

MAPPER Session and Instructions A-6

Obtaining Printout A-8

SG 3605 xiii
Relational Concepts and SQL Programming

Compiling and Executing a Program in IPF A-9

Compiling and Executing a Program using ECL A-10

Use Runstreams A-10
ACOB program A-10
UCOB program A-10

Saving and Printing Compiler Listings A-11

Basic IPF Commands A-12

Command Line Commands A-12
Text Line Commands A-12

Appendix B: FORTRAN Examples

SQL Commands in ASCII FORTRAN B-3

Sample ASCII FORTRAN Retrieval Program B-4

Example With GETERROR (FTN) B-6

Appendix C: ACOB Skeleton

Appendix D: ACOB Examples

Appendix E: UCOB Skeleton

Appendix F: UCOB Examples

xiv SG 3605
 Relational Concepts and SQL Programming

Course Description

Relational Concepts and SQL Programming

Objectives:
To prepare the student to write, test, and debug SQL programs.

Audience:
Application programmers, systems analysts, systems designers, and support personnel

Prerequisites:
Extensive COBOL programming experience in the 1100/2200 environment

Key Topics:
• Introduction to RDMS
• SQL command set
• Embedding SQL in programs
• Error handling and debugging
• IPF SQL
• MAPPER Relational Interface (MRI)

Relational Concepts and SQL Programming

EL3605 5 DAYS

SG 3605 XV
Relational Concepts and SQL Programming

About This Course

How to Use This Document

This student guide supports the material the instructor presents. A complete listing of
referenced product information is provided under the Referenced Product Information
heading.

There are written and/or laboratory practices at the end of most modules to help you
understand the material presented. Two modules, modules 3 and 9, contain laboratory
exercises in the middle of the module.

Learning Environment
This course is designed to provide you with the background and skills to manipulate an
RDMS 1100 database. The concepts of relational database are introduced from a
programming viewpoint. The SQL command set is used to enable you to access and
manipulate relational data through COBOL host language programs. The IPF SQL and
MAPPER Relational Interface are introduced as other ways to access RDMS 1100
databases. Lab exercises provide you with the opportunity to practice coding skills in
relational data creation, retrieval, and updating, and use of alternative interfaces to
RDMS 1100.

Trademark Information

• MAPPER is a registered trademark of Unisys Corporation

• LINC and ALLY are registered trademarks of Unisys Corporation

xvi SG 3605
 Relational Concepts and SQL Programming

Referenced Product Information

OS 1100 UDS Relational Data Management System (UDS RDMS 1100) and IPF SQL
Interface End Use Guide 7831 0778-000

OS 1100 UDS Relational Data Management System (UDS RDMS 1100) SQL
Programming Reference Manual 7830 8160-002

OS 1100 UDS Relational Data Management System (UDS RDMS 1100) Administration
Guide 7831 0760-002

OS 1100 Universal Compiling System (UCS) COBOL Programming Reference Manual

Volume 2: Compiler and System Interface 7831 0455-000

OS 1100 Universal Compiling System (UCS) Application Development Programming

Guide, Volume 1 7831 4077-001

MAPPER Relational Interface (MRI) Relational Database Interface (RDI) Operations

Guide 7831 9795-000

Note: Consult the Series 1100 and 2200 Systems Product Information Library
Directory, if needed, for specific levels of these documents and other Unisys
documentation. The proper level of a document is needed to match release of
installed software for your system.

SG 3605 xvii
Relational Concepts and SQL Programming

Agenda
Day 1
Introduction

Retrieving Data with the SELECT Statement

Using SQL in Application Programs

Day 2
More WHERE Clause Operators

The IPF SQL Interface

The MAPPER Relational Interface

Day 3
Grouping Functions

Joining Tables

More Retrieval Considerations

Day 4
Data Manipulation

Keys and Updating

Obtaining Table Definition Information

Day 5
Views

More About Threads and UDS

Getting Physical - Tables and Storage Areas

xviii SG 3605
1
Introduction
 Introduction

Module 1
Introduction

Module Objectives 1-3

Benefit Statement 1-3

Materials 1-3

Relational Database 1-4

Operations on Tables 1-5

RESTRICT Operation 1-6
PROJECT Operation 1-7
JOIN Operation 1-8

Advantages of the Relational Model 1-9

What is RDMS 11007 1-10

Interfaces to RDMS I 1-11

Interfaces to RDMS II 1-12

The Structured Query Language (SQL) 1-13

Exercise 1-1 1-14

SG 3605 1-1
 Introduction

Module Objectives

Upon completion of this module, you should be able to

• Identify some of the basic features of a relational database management system.

- Describe the uses and advantages.

- Correctly use the terms to describe a table, primary keys, tuples, and the
operations that can be performed on tables.

- Identify three interfaces into RDMS databases.

Benefit Statement

It is important that you have a few, fundamental concepts and terms in order to visualize
and work with a relational database. You will realize the several options you have for
accessing the database and understand how these options are implemented.

Materials

• OS 1100 UDS Relational Data Management System (UDS RDMS 1100) and IPF SQL
Interface End Use Guide

SG 3605 1-3
Introduction

Relational Database

A database in which all the data is represented by tables. Table 1-1 illustrates a table.

Table 1 -1. National League Winners

YEAR CLUB WON LOST MANAGER

1876 CHICAGO 52 14 SPAULDING, ALBERT

1877 BOSTON 31 17 WRIGHT, HARRY
1878 BOSTON 41 19 WRIGHT, HARRY
1879 PROVIDENCE 55 23 WRIGHT, GEORGE
1880 CHICAGO 67 17 ANSON, ADRIAN
1881 CHICAGO 56 28 ANSON, ADRIAN
1882 CHICAGO 55 29 ANSON, ADRIAN
1883 BOSTON 63 35 MORRILL, JOHN
1884 PROVIDENCE 84 28 BANCROFT, FRANK
1885 CHICAGO 87 25 ANSON, ADRIAN
1886 CHICAGO 90 Si— ANSON, ADRIAN

• Relational database terms

Schema

- Entity

- Table (relation)

Column (attribute)

- Row (tuple)

- Primary key

Data value

Secondary key

1-4 SG 3605
 Introduction

Operations on Tables

There are three basic types of operations that can be performed on tables:

• RESTRICT - Choose rows from a table

• PROJECT - Choose columns from a table

• JOIN Combine two or more tables that have a cobimn in common

These are the formal names for operations in the database literature. They are not the
commands used in SQL to perform the operations, but are listed to illustrate the kind of
processing that is possible.

The result of any of the above operations is always a conceptual table.

These three operations may be combined to produce the exact data desired.

SG 3605 1-5
Introduction

RESTRICT Operation
The RESTRICT operation extracts a subset of the rows of an existing table. The desired
rows are selected by stating the conditions that the data values in specified columns must
satisfy. RESTRICT creates a horizontal subset of the existing table. From Table 1-2 we
extract only the winners from 1880 to 1885 creating Table 1-3.

Table 1-2. National League Winners

YEAR CLUB WON LOST MANAGER

1876 CHICAGO 52 14 SPAULDING, ALBERT

1877 BOSTON 31 17 WRIGHT, HARRY
1878 BOSTON 41 19 WRIGHT, HARRY
1879 PROVIDENCE 55 23 WRIGHT, GEORGE
1880 CHICAGO 67 17 ANSON, ADRIAN
1881 CHICAGO 56 28 ANSON, ADRIAN
1882 CHICAGO 55 29 ANSON, ADRIAN
1883 BOSTON 63 35 MORRILL, JOHN
1884 PROVIDENCE 84 28 BANCROFT, FRANK
1885 CHICAGO 87 25 ANSON, ADRIAN
1886 CHICAGO 90 34 ANSON, ADRIAN

Table 1-3. Winners from 1880 to 1885.

YEAR CLUB WON LOST MANAGER

1880 CHICAGO 67 17 ANSON, ADRIAN

1881 CHICAGO 56 28 ANSON, ADRIAN
1882 CHICAGO 55 29 ANSON, ADRIAN
1883 BOSTON 63 35 MORRILL, JOHN
1884 PROVIDENCE 84 28 BANCROFT, FRANK
1885 CHICAGO 87 25 ANSON, ADRIAN

1-6 SG 3605
 Introduction

PROJECT Operation
The PROJECT operation extracts a subset of the columns of an existing table. The
desired columns are selected by stating the column names.

PROJECT creates a vertical subset of the existing table. Table 1-5 is produced from Table
1-4 by projecting the YEAR, CLUB and MANAGER columns.

Table 1-4. National League Winners

YEAR CLUB WON LOST MANAGER

1880 CHICAGO 67 17 ANSON, ADRIAN

1881 CHICAGO 56 28 ANSON, ADRIAN
1882 CHICAGO 55 29 ANSON, ADRIAN
1883 BOSTON 63 35 MORRILL, JOHN
1884 PROVIDENCE 84 28 BANCROFT, FRANK
1885 CHICAGO 87 25 ANSON, ADRIAN
1886 CHICAGO 90 34 ANSON, ADRIAN

Table 1-5. Projected Columns

YEAR CLUB MANAGER

1880 CHICAGO ANSON, ADRIAN

1881 CHICAGO ANSON, ADRIAN
1882 CHICAGO ANSON, ADRIAN
1883 BOSTON MORRILL, JOHN
1884 PROVIDENCE BANCROFT, FRANK
1885 CHICAGO ANSON, ADRIAN
1886 CHICAGO ANSON, ADRIAN

SG 3605 1-7
Introduction

JOIN Operation
=
The JOIN operation creates a virtual table by combining the columns from two or more
tables. The tables must each have a column the values of which come from the same
domain (have the same data type). The user specifies which columns to use for the JOIN
and how to compare them.

Table 1-6. Club Winners Table 1-7. Team Names

YEAR CLUB WINNER CLUB NICKNAME

1880 CHICAGO CHICAGO WHITE STOCKINGS

1881 CHICAGO BOSTON RED STOCKINGS
1882 CHICAGO PROVIDENCE GRAYS
1883 BOSTON
1884 PROVIDENCE
1885 CHICAGO
1886 CHICAGO

Tables 1-6 and 1-7 are joined to display for each year the year, club, and nickname in
Table 1-8.

Table 1-8. Team Winners for 1880 - 1886

YEAR CLUB NICKNAME

1880 CHICAGO WHITE STOCKINGS

1881 CHICAGO WHITE STOCKINGS
1882 CHICAGO WHITE STOCKINGS
1883 BOSTON RED STOCKINGS
1884 PROVIDENCE GRAYS
1885 CHICAGO WHITE STOCKINGS
1886 CHICAGO WHITE STOCKINGS

1-8 SG 3605
 Introduction

Advantages of the Relational Model

• All data is viewed as two-dimensional tables

- Simplicity: easily understood by users with no programming experience

- Uniformity: all data is stored in tables and output is in table format

• No need to know how the data is stored

- Access is through data values

- Query language (SQL) is simple

• Relationships are implicit in the data

- Allows easy handling of unanticipated uses of the data or ad hoc queries

- Enables addition or modification of related data

SG 3605 1-9
Introduction

What is RDMS 1100?

User Intertecs

Universal
Data System
RDMS

UREP )1. LIDS Control

Figure 1-1. RDMS 1100

Relational Data Management System (RDMS)

The database manager for relational data.

Unisys Repository Manager (UREP)

The Unisys database management software that organizes data for user applications in
an integrated information management system and in individual files. UREP also
maintains the repository.

Universal Data System (UDS) Control

On-line data manager for all three UDS data models. Provides control for common
functions such as 1/0, locking, queuing, memory management, and recovery.

Universal Data System (UDS)

A suite of Unisys software products for database management, data processing, and
database application development. Together with other software components and database
software products, UDS provides an integrated environment for control, maintenance, and
recovery of user databases.

1-10 SG 3605
 Introduction

Interfaces to RDMS I

IPF COBOL FORTRAN MRI

SQL Program Program Run

RDMS

Figure 1-2. RDMS Interfaces

• Figure 1-2 describes the interfaces discussed and illustrated in this course.

• RDMS also provides these other interfaces:

- Pascal and C programs from the Universal Compiling System (UCS)

- Query Language Processor (QLP)

- Logic and Information Network Compiler II (UNC II)

- ALLY Software

- SQL*Plus 1100

- SQL*Report 1100

- SQL*OCI 1100

SG 3605 1-11
Introduction

Interfaces to RDMS II

Application programs

— SQL commands can be placed in ASCII COBOL and ASCII FORTRAN programs

— SQL commands can be placed in Universal Compiling System (UCS) programs

written in COBOL, FORTRAN, C and Pascal

• IPF SQL

— Interactive access to RDMS databases using SQL commands

— SQL files can be created and run in IPF to execute a series of SQL commands

• MAPPER Relational Interface (MRI)

- Interactive menu-driven interface to construct SQL commands

- Data selected is returned as a MAPPER report

- Report can be further manipulated with MAPPER functions

- Run statements also available for embedding in MAPPER runs

1-12 SG 3605
 Introduction

The Structured Query Language (SQL)

• Industry standard relational programming language

"English-like" non-procedural language

- Users do not specify how to access the data

• SQL can be interactive or programmatic

— Embedded in COBOL, FORTRAN, Pascal, or C programs, or MAPPER runs

— Ad hoc queries through IPF SQL or MAPPER MRI

• Language statements in four categories

- Queries = SELECT...FROM...

- Data Manipulation = DELETE, INSERT, UPDATE

Data Definition = CREATE, ALTER, DROP

- Data Control = GRANT, REVOKE

SG 3605 1-13
Introduction

Exercise 1-1

Provide answers or definitions as requested.

1. List at least three advantages of using a relational database management system.

2. Identify three interfaces to RDMS 1100 data.

3. Define the following relational database terms:

a. relation

b. tuple

c. attribute

d. primary key

1-14 SG 3605
2
Retrieving Data with
the SELECT Statement
 Retrieving Data with the SELECT Statement

Module 2
Retrieving Data with the SELECT Statement

Module Objectives 2-3

Benefit Statement 2-3

Materials 2-3

SQL Queries 2-4

Sample Tables for Examples and Exercises 2-5

SELECT Command 2-6

Selecting Columns 2-7
DISTINCT Values 2-8
The WHERE Clause 2-9
WHERE Clause Operators 2-10
WHERE Clause Usage 2-11
The ORDER BY Clause 2-12
Arithmetic Expressions 2-13
Order of Precedence 2-14

Exercise 2-1 2-15

SG 3605 2-1
 Retrieving Data with the SELECT Statement

Module Objectives

Upon completion of this module, you should be able to

• Formulate simple SELECT statements to retrieve data from a database.

- Including basic use of the WHERE clause.

- Using arithmetic operators to retrieve data derived from numeric columns of a

table.

Benefit Statement

The most basic SQL command is the SELECT query specification. You will need this
command to retrieve information from your database. It will be used alone or in
combination with other commands. A few simple SELECT command clauses will give you
a great deal of power for your database accesses.

Materials

• OS 1100 UDS Relational Data Management System (UDS RDMS 1100) and IPF SQL
Interface End Use Guide

• OS 1100 UDS Relational Data Management System (UDS RDMS 1100) SQL
Programming Reference Manual

SG 3605 2-3
Retrieving Data with the SELECT Statement

SQL Queries

The SELECT statement is used to retrieve data from a table in the database.

• Variations of the SELECT statement are used to perform all table operations

- Restrict
- Project
Join

• Can SELECT
- Entire table
Some columns

Some rows

• Can display columns in any order desired

• Order of rows displayed is not (logically) significant

2-4 SG 3605
 Retrieving Data with the SELECT Statement

Sample Tables for Examples and Exercises

Table 2-1. EMP

EMPNO ENAME JOB MGR DNO HIREDATE SAL COMM

5010 FOSTER SALESREP 5234 200 860714 2700 400
5146 BROWN CLERK 5234 200 871011 1000
5234 WOODWORTH MANAGER 5784 200 790219 3250
5237 ROCKWELL ACCOUNTANT 5743 100 840618 2175
5437 MART IN SALESREP 5234 200 870326 2500 300
5469 ADAMS SALESREP 5234 200 811102 2250 500
5630 GLASS ACCOUNTANT 5743 100 871220 1850
5702 TURNER CLERK 5743 100 890515 900
5743 LAWSON MANAGER 5784 100 830510 3400
5765 JOHNSON CLERK 5984 400 891221 975
5784 WILLIAMS PRESIDENT 300 781015 5200
5896 SMITH CLERK 5984 400 880913 1050
5942 FORD SALESREP 5234 200 890316 2375 0
5984 HYDE MANAGER 5784 400 850901 2900

Table 2-2. DEPT Table 2-3. SAL GRADE

DNO DNAME LOC GRADE LOSAL HISAL

100 ACCOUNTING ATLANTA 1 800 1599
200 SALES NEW YORK 2 1600 2199
300 MARKETING DETROIT 3 2200 2799
400 DISTRIBUTION PRINCETON 4 2800 3699
5 3700 5999

Table 2-4. AUX_EMP

EMPNO GRADE SEX MAR STAT DEPENDENTS

5010 3 F S 0
5146 1 M M 2
5234 4 M S 0
5237 2 m S 0
5437 3 F M 1
5469 3 M M 2
5630 2 M M 3
5702 1 F M 0
5743 4 M M 3
5765 1 M S 0
5784 5 F S 0
5896 1 M S 0
5942 3 F M 2
5984 4 F S 0

SG 3605 2-5
Retrieving Data with the SELECT Statement

SELECT Command

SELECT [ DISTINCT ] column1, column2, column3

FROM tablename

• Select-list lists columns to return in desired order

• To display all columns replace select-list with *

Example:
SELECT *
FROM DEPT

DNO DNAME LOC

100 ACCOUNTING ATLANTA

200 SALES NEW YORK
300 MARKETING DETROIT
400 DISTRIBUTION PRINCETON

2-6 SG 3605
 Retrieving Data with the SELECT Statement

Selecting Columns

• Collin-ma are displayed in the order requested in the select-list

Example: Display the department name, location, and department number, in that order,
of all departments in the DEPT table.

SELECT DNAME, LOC, DNO

FROM DEPT

DNAME LOC DNO

ACCOUNT ING ATLANTA " 100

SALES NEW YORK 200
MARKET ING DETROIT 300
DISTRIBUTION PRINCETON 400

• Not all columns need be displayed

Example: Display the department name and department number columns from the
DEPT table.

SELECT DNAME, DNO

FROM DEPT

DNAME DNO

ACCOUNTING 100
SALES 200
MARKETING 300
DISTRIBUTION 400

SG 3605 2-7
Retrieving Data with the SELECT Statement

DISTINCT Values CTS

• The DISTINCT keyword is used to suppress duplicate values in a column selected for
output.

Example: Display the different types of jobs in all the departments without repetition.

SELECT DISTINCT JOB

FROM EMP

JOB

ACCOUNTANT
CLERK
MANAGER
PRESIDENT
SALESREP

• If more than one column is selected, DISTINCT selects unique combinations of values
in those columns

Example: Display all different jobs existing in each department

SELECT DISTINCT DNO, JOB

FROM EMP

DNO JOB

100 ACCOUNTANT
100 CLERK
100 MANAGER
200 CLERK
200 MANAGER
200 SALESREP
300 PRESIDENT
400 CLERK
400 MANAGER

2-8 SG 3605
 Retrieving Data with the SELECT Statement

The WHERE Clause

SELECT select-list
FROM tablename
WH ERE condition

• Used to specify which rows to retrieve

• Specifies condition(s) which the column value must satisfy

Example: Display the employee information for all employees named Hyde.

SELECT *
FROM EMP
WHERE ENAME = 'HYDE'

EMP NO ENAME JOB MGR DNO HIREDATE SAL COMM

5984 HYDE MANAGER 5784 400 850901 2900

Example: Display the name and position of everyone in Department 200.

SELECT ENAME, JOB

FROM EMP
WHERE DNO = 200

ENAME JOB

FOSTER SALESREP
BROWN CLERK
WOODWORTH MANAGER
MARTIN SALESREP
ADAMS SALE SREP
FORD SALESREP

SG 3605 2-9
Retrieving Data with the SELECT Statement

WHERE Clause Operators

These binary operators are used to compare two values, either numeric or literal.

• = Equal

— WHERE JOB = 'MANAGER'

• <> Not equal

— WHERE DNO <> 400

• > Greater than

— WHERE COMM > SAL

• < Less than

— WHERE COMM < SAL

• >= Greater than or equal to

— WHERE HIREDATE >= 820101

• <= Less than or equal to

— WHERE HIREDATE <= 850331

2-10 SG 3605
 Retrieving Data with the SELECT Statement

WHERE Clause Usage

Columns named in the WHERE clause must belong to the table named in the FROM
clause.

• Columns named in the WHERE clause need not be in the select-list

— The values tested in the condition do not need to be displayed in the resulting
table

• Character string data must be enclosed in single quotes or apostrophes

• Case is significant in character strings.

— 'HYDE' is not equal to 'Hyde'

SG 3605 2-11
Retrieving Data with the SELECT Statement

The ORDER BY Clause

SELECT select-list
FROM tablename
WHERE condition
ORDER BY sort-specification-list

• Controls how rows returned are displayed

• Sort specification format

[ unsigned-integer I column-name ] [ ASC I DESC

• Integer refers to position in the select-list

• Default order is ascending ASC

Example: Display the employee information for department 400 in alphabetical order by
employee name

SELECT EMPNO, ENAME, JOB, MGR, DNO, HIREDATE, SAL

FROM EMP
WHERE DNO = 400
ORDER BY ENAME ASC

EMPNO ENAME JOB MGR DNO HIREDATE SAL

----
5984 HYDE MANAGER 5784 400 850901 2900
5765 JOHNSON CLERK 5984 400 891221 975
5896 SMITH CLERK 5984 400 880913 1050

2-12 SG 3605
 Retrieving Data with the SELECT Statement

Arithmetic Expressions
Arithmetic expressions can be used to display data that is derived from values in columns
of the select-list. An expression is composed of column names and constant numeric
values connected by an arithmetic operator.

Table 2-5. Arithmetic Operators

+ ADD
- SUBTRACT
* MULTIPLY
/ DIVIDE

Example: Display the names and yearly salaries of all employees in Department 400.

SELECT ENAME, SAL * 12

FROM EMP
WHERE DNO = 400

ENAME

JOHNSON 11700
SMITH 12600
HYDE 34800

SG 3605 2-13
Retrieving Data with the SELECT Statement

Order of Precedence
In evaluating expressions, the order of precedence of operators is as follows:

Unary operators positive (+) and negative (-)

• Multiplication (*) and division

▪ Addition (+) and subtraction (-)

Equal precedence operations, like addition and subtraction, are performed in order, left to
right as encountered.
Parentheses are used to clarify or change the order of operations. Nested parentheses are
evaluated beginning with the innermost and working outward.

Example:
SAL + COMM * 12 - 200 is evaluated as

SAL + COMM * 12 - 200.

1---
2
3-

Parentheses can change the order in which expressions are evaluated.

(SAL + COMM) * 12 - 200 is evaluated as

(SAL + COMM) * 12 - 200.

1
2 ---
3

2-14 SG 3605
 Retrieving Data with the SELECT Statement

Exercise 2-1

Write the SQL statement necessary to access either the EMP or DEPT table to retrieve
the following information. Indicate which records will be displayed.

1. Display the department name and department number for the department in
Atlanta.
CAI)

2. Display the names of all employees whose total compensation is more than $2600
per month.
„„•

3. Display the name, monthly salary, daily salary, and hourly salary for all
employees. Assume the SAL column is monthly salary and that there are
22 working days in a month and 8 hours in a working day.

4. Display the names and total compensation of all employees whose

commissions are more than 20% of their salaries. Have the names
displayed in order by within the departments.
•

SG 3605 2-15
3
Using SQL in
Application Programs
 Using SQL in Application Programs

Module 3
Using SQL in Application Programs

Module Objectives 3-3

Benefit Statement 3-3

Materials 3-3

Using SQL Commands in 3GL Programs 3-4

Accessing RDMS 1100 Data 3-4
3GL Interfaces 3-4
UDS Environment Commands 3-4

SQL and Unisys Extensions 3-5

Summary of Selected SQL Commands 3-6

Comparing Embedded and Interpretive SQL 3-8

Source 3-9
Compilation 3-9
Collection or Link 3-9
Execution 3-9

SQL Commands in COBOL Programs 3-10

Interpretive SQL 3-10
Embedded SQL 3-12

Command Building Blocks 3-15

Constants 3-15
Naming Database Objects 3-15

RDMS 1100/2200 Data Types 3-16

Character Data Types 3-16
Numerical Data Types 3-16
RDMS and 3GL Data Type Equivalencies 3-18

RDMS 1100/2200 Terminology 3-19

Cursor 3-19
Currency 3-19
Application Group 3-19
Thread 3-19

Program Flow 3-19

BEGIN THREAD Command 3-21

SG 3605 3-1
 Using SQL in Application Programs

DECLARE CURSOR Command 3-22

OPEN CURSOR Command 3-23

FETCH Command 3-24

Passing Information Between a Program and RDMS 3-25

Host Program Variables 3-25
Placeholder Variables in ISQL 3-25
Embedded Variables in ESQL 3-25
Parameter Passing Limitations 3-25
ISQL - Placeholder and Host Program Variables 3-26

END THREAD Command 3-27

USE DEFAULT QUALIFIER Command 3-28

Exercise 3-1 3-29

Lab 3-1 3-31

Error Handling Choices 3-33

RDMS Error Handling 3-34
GETERROR Command 3-35
Example Using GETERROR (ACOS) 3-36
ESQL Error Handling 3-37
Data and Commands for ESQL Error Handling 3-38
WHENEVER Command 3-39
ESQL Data Area Declarations 3-40
Example Using GETERROR (UCOB) 3-41

Lab 3-2 3-42

3-2 SG 3605
 Using SQL in Application Programs

Module Objectives
Upon completion of this module, you should be able to

• Incorporate SQL commands in a host programming language.

- Describe the flow of processing in a host program that accesses RDMS 1100.

- Use the BEGIN THREAD and END THREAD commands to establish and
terminate a thread with UDS Control.

- Use SQL commands in a program to retrieve data from an RDMS database.

- Use host or embedded (ESQL) variables to pass parameters between RDMS and
host programs.

- Include error code in a host program to handle RDMS errors.

- Compile and run a program containing SQL commands.

Benefit Statement
The SELECT command you have already seen is standard SQL. If you use an interactive
interface like IPF SQL, you would enter the command just as you have seen it.

However, if you are not performing ad hoc queries and need to further manipulate the
.data after it has been retrieved, your SQL commands must be embedded in a third-
generation language (3GL) program. This module tells you how to incorporate SQL into
COBOL or FORTRAN programs in the 1100/2200 environment. As a programmer, you
need to be familiar with the interfaces between RDMS and UDS and your 3GL program.

Materials
• OS 1100 Universal Compiling System (UCS) COBOL Programming Reference Manual
Volume 2: Compiler and System Interface

• OS 1100 Universal Compiling System (UCS) Application Development Programming

Guide, Volume 1

• OS 1100 UDS Relational Data Management System (UDS RDMS 1100) SQL
Programming Reference Manual

SG 3605 3-3
Using SQL in Application Programs

Using SQL Commands in 3GL Programs

Accessing RDMS 1100 Data

RDMS 1100 data can be accessed using standard SQL syntax in your program. RDMS
1100 also supports some SQL extensions and alternative formats that are not part of the
ANSI standard. The syntax and commands that are nonstandard will be made clear to
you in the text. If you are concerned about portability of your program, use only the
standard syntax.

You have seen that a SELECT command returns a whole table or a subset of a table. A
third-generation language (3GL) program typically deals with only one record or row of
the returned table at a time. How to transfer one record of the returned table to the
program, and how to code program variables that reference them are important issues.

3GL Interfaces
Unisys provides two interfaces for 3GL programming languages - an interpretive SQL and
embedded SQL or ESQL.

Interpretive SQL incorporates the SQL statement as a string literal passed as an

argument to RDMS on a procedure call. This interface was originally available from
ASCII COBOL and ASCII FORTRAN, and made available to the languages in the
Universal Compiling System (UCS), namely UCS COBOL, UCS FORTRAN, UCS C, and
UCS Pascal. RDMS must interpret each SQL statement each time it is encountered
during program execution. Interpretive SQL will be abbreviated ISQL in this student
guide whenever an ISQL - ESQL distinction is important.

The Embedded SQL interface, which is only available from UCS COBOL, fully conforms
to the SQL standard and provides some extensions as well. Each SQL command begins
with an embedded SQL prefix, EXEC SQL, and ends with an embedded SQL terminator,
END-EXEC. The embedded SQL commands are fully compiled along with the host
program. Execution of ESQL commands will be significantly faster than their interpreted
counterparts. This is particularly noticeable in program loops. Interpretive and embedded
SQL commands can be used in the same program.

Several other differences exist between the two implementations. ASCII COBOL and UCS
COBOL will be used to illustrate these differences.

UDS Environment Commands

Another important issue is how to incorporate your program into the UDS environment so
that it takes advantage of the UDS recovery and locking features. Specific commands
needed to accomplish this are also shown.

3-4 SG 3605
 Using SQL in Application Programs

SQL and Unisys Extensions

• Accessing RDMS 1100 data

- ANSI standard SQL

- Unisys extensions

- One record at a time

• 3GL interfaces

- Interpretive SQL

Embedded SQL

• UDS environment commands

- Provide recovery

- Set up UDS environment

- Use Unisys extensions

SG 3605 3-5
Using SQL in Application Programs

Summary of Selected SQL Commands

Table 3-1 groups the standard SQL commands that will be covered in this course. Unless
noted by the asterisk in Table 3-1, RDMS supports the SQL 89 form of the command and,
in many cases, offers one or more extensions to it. Table 3-1 is derived from the SB4R4
system release, which is release 5R3 of RDMS.

Refer to the OS 1100 UDS Relational Data Management System (UDS RDMS 1100) SQL
Programming Reference Manual for more the details on the supported extensions for the
above commands.

Table 3-1. Standard SQL Commands

Function Command

Data Definition CREATE TABLE *

Data Retrieval DECLARE CURSOR
OPEN
CLOSE
FETCH
SELECT Single Row
Data Creation/ INSERT
Manipulation DELETE
COMMIT
ROLLBACK
BEGIN DECLARE (ESQL)
END DECLARE (ESQL)
UPDATE Searched
UPDATE Positioned

Miscellaneous GRANT
WHENEVER (ESQL) *

* RDMS does not support the SQL 89 form of this command and offers
no extensions to it

3-6 SG 3605
 Using SQL in Application Programs

Table 3-2 lists the SQL extensions available from Unisys that will be covered in this
course.

Table 3-2. Common Unisys Extensions for the 1100/2200 Environment

Function Command

Data Definition ALTER TABLE

DROP TABLE
CREATE INDEX
DROP INDEX
CREATE VIEW
DROP VIEW
Data Retrieval DROP CURSOR **
FETCH (next/prior/first/last/current)
FETCH NEXT n
LOCATE
Data Locking LOCK
UNLOCK
Thread Control BEGIN THREAD
END THREAD
Miscellaneous GETERROR
USE DEFAULT

** Not allowed in embedded SQL (ESQL)

SG 3605 3-7
Using SQL in Application Programs

Comparing Embedded and Interpretive SQL

ESQL ISQL
Source: UCS COBOL Source: ASCII COBOL

EXEC SOL ENTER MASM 'ACOB$RDMR`

USING 'sql command'
END-OlEcemmand Orr-Stat
aux-info

@UCOB @ACOB

Dynamic/
Static @MAP
Linking

@XQT @XQT

RDMS

Figure 3-1. ESQL/ISQL Comparison

3-8 SG 3605
 Using SQL in Application Programs

Source

- Place embedded UCOB SQL (ESQL) commands in packets delimited by EXEC SQL
and END-EXEC.

- Format interpretive SQL commands (ISQL) according to the host language

conventions

- ASCII or UCS COBOL - ENTER MASM 'ACOB$RDMR' commands

- ASCII or UCS FORTRAN - CALL F$RDMR commands

UCS Pascal - procedure call statement to F$RDMR

- UCS C - include the system header file rsa.h, then call the rsa function

Compilation

- Embedded commands are fully compiled

- Analyze SQL syntax

- Create an internal form of the command

- Resolve table and column names (tables must be already created)

- Optimize access path

- Interpreted commands are not compiled (no prior creation of tables needed)

Collection or Link

- Resolve references between user elements and support libraries

- Create an executable program structure

Execution

- Embedded SQL commands executed to access database

- Interpretive SQL commands are compiled, then executed if no syntax errors exist

SG 3605 3-9
Using SQL in Application Programs

SQL Commands in COBOL Programs

Interpretive SQL
Format 1

ENTER MASM 'ACOB$RDMR' USING

'SQL-COMMAND •I 'I
err-stat,
aux-info.

Format 2

MOVE sql-corn-string-variable TO rcom.

ENTER MASM 'ACOB$RDMR' USING rcom, err-stat, aux-info.

Syntax Notes

• The SQL command or the contents of SQL-COM-STRING-VARIABLE must end with

semicolon (;).

• The item err-stat defines an error status returned by RDMS for an SQL command;
normal completion is 0000.

• The item aux-info contains useful information about the error status.

3-10 SG 3605
 Using SQL in Application Programs

ISM Example

DATA DIVISION.
WORKING-STORAGE SECTION:

01 rcom.
05 rlin PIC X(40) OCCURS 10 TIMES.
01 err-stat PIC 9(4).
01 aux-info PIC S1(36).

•
PROCEDURE DIVISION.

ENTER MASM 'ACOB$RDMR' USING

'BEGIN THREAD FOR udssrc READ
err-stat,
aux-info.

MOVE 'DECLARE employee CURSOR SELECT * FROM emp

TO rcom.
ENTER MASM 'ACOB$RDMR' USING rcom, err-stat, aux-info.

SG 3605 3-11
Using SQL in Application Programs

Embedded SQL
Format 1

EXEC SQL
SQL-COMMAND
END-EXEC.

Format 2

SQL-COMMAND.

Syntax Notes

• Format 1

- Required by SQL standard

- Delimit packets for the compiler with EXEC SQL and END-EXEC

- Insert only one SQL command into a packet

• Format 2

- For five commands that do not require the packet delimiters (packet is optional)

- BEGIN THREAD

- END THREAD

- COMMIT

- ROLLBACK

- GETERROR

- Command must end with a period

- Command cannot be part of a larger command structure, for example an IF

statement

3-12 SG 3605
 Using SQL in Application Programs

ESQL Command Scoping

• Five commands are acted upon during compilation

- DECLARE CURSOR

USE DEFAULT

- WHENEVER

- DEBUG (static format only)

- SET STATISTICS

• Must be placed physically ahead of other commands they affect in the source program

Note: This course will address static ESQL only. Dynamic ESQL provides a subset of
SQL commands, typically involving a database update, that are not processed
at compile time, but at execution time. Dynamic ESQL is the functional
equivalent of the ENTER MASM interpretive interface for COBOL. A program
can contain a mixture of both static ESQL and interpretive SQL; this fact may
be important in a migration to ESQL from ISQL.

SG 3605 3-13
Using SQL in Application Programs

ESQL Example

Notice the SQL commands in the Procedure Division illustrated below that do or do not
require ESQL packets. Part of required Data Division is illustrated; it will be more fully
discussed with error handling in this module.

DATA DIVISION.
WORKING-STORAGE SECTION.

01 SQLCODE PIC S9(9) USAGE BINARY.

EXEC SQL. BEGIN DECLARE SECTION END-EXEC.
01 RDMCA.
05 ERROR-STATUS PIC 9(4).
05 AUX-INFO PIC S9(9) USAGE BINARY.
EXEC SQL END DECLARE SECTION END-EXEC.

PROCEDURE DIVISION.

EXEC SQL
WHENEVER SQLERROR GO TO :rdms-err-para
END-EXEC.

BEGIN THREAD FOR udssrc READ.

EXEC SQL
DECLARE employee CURSOR SELECT * FROM emp
END-EXEC.

rdms-err-para.

3-14 SG 3605
 Using SQL in Application Programs

Command Building Blocks

Constants

String Literals

A string literal is a character sequence surrounded by string delimiters. Single quotes

(apostrophes) are string delimiters.

ENTER MASM 'ACOB$RDMR' USING 'SQL Command' ... uses a string literal to pass the
command to RDMS.

Numeric Literals

Numeric literals are either integers or real numbers. Real numbers contain a decimal
point, but integers do not. You can write them with or without the leading sign. Numeric
literals cannot be longer than 21 digits. This limit does not include either the sign or the
decimal point.

NULL Values

A data item that has no assigned value has a NULL value. NULL values are distinct
from zero values or spaces. NULLs are used instead of inserting dummy values that may
be mistaken for assigned values. NULL values may not be used in arithmetic or Boolean
expressions. There are special functions that test for NULL values.

Naming Database Objects

1. Names may contain from 1 to 30 characters.

2. Names may use characters A-Z, a-z, 0-9, and _(underscore).

3. Names must begin with a letter, and cannot end with an underscore.

4. Names are not case sensitive.

5. The name may not be a reserved word.

6. The above rules can be ignored if the name is enclosed in double quotes.

Table 3-3 illustrates several valid and invalid names for database objects.

SG 3605 3-15
Using SQL in Application Programs

Table 3-3. Examples of Database Object Names

Valid Invalid

sales_reps sales-reps
PURCHASE_PRICE PURCHASE*PRICE
"TABLE" TABLE

RDMS 1100/2200 Data Types

Every literal or variable has a data type. A data type associates a fixed set of properties
with a value. It also restricts the values the data item may contain

A value in the database takes on the data type of the column to which it belongs. When a
table is created or modified, the column is assigned a data type. When a query is
executed and values are returned to your program, the program variables into which the
values will go must correspond to the data type of the database item.

RDMS supports two character and seven numerical data types.

Character Data Types

The character data types and their definition formats are

CHARACTER (size)

NCHARACTER (size)

CHARACTER data types contain ASCII or any other 9-bit data. Character(5) specifies a
five-position character string. A corresponding COBOL program variable should be
defined as PIC X(5).

NCHARACTER contains 16-bit data, such as Kanji characters.

Numerical Data Types

There are several data types that hold numerical information. Numerical data types can
be either exact or approximate.

3-16 SG 3605
 Using SQL in Application Programs

Exact Numerical Data Types

The exact numerical data types and their definition formats are:

DECIMAL [(sign-and-digits-precisionLscaleN

NUMERIC [digits-precisionkscale)]

INTEGER

SMALLINT

The value of sign-and-digits-precision specifies the total number of digits in the column
including the leading sign. This must be an integer between 2 and 22, inclusive. For
example, DECIMAL(6.2) allows 6 positions, 2 of which are decimal. A corresponding
COBOL program variable would be defined as PIC S999V99. Decimal(4) should be
declared as PIC S9(3).

The digits-precision is the number of digits in the column and must be no larger than 11.

The scale is a non-negative integer that indicates the number of digits for the fractional
portion of the column This can be no larger than digits precision - 1 for NUMERIC and
no larger than sign-and-digits-precision -1 for DECIMAL.

A period can be used instead of a comma between digits precision and scale.

Approximate Numerical Data Types

The approximate numerical data types and their formats are:

REAL

FLOAT [(binary-precision)]

DOUBLE PRECISION

The value of binary precision must be a positive integer less than 60. It defines the
minimum binary precision in bits.

Consult the OS 1100 UDS Relational Data Management System (UDS RDMS 1100) SQL
Programming Reference Manual for more information on the numerical data types.

SG 3605 3-17
Using SQL in Application Programs

RDMS and 3GL Data Type Equivalencies

Both RDMS and your host language define data types. You should use equivalent data
types for corresponding data items in order to avoid unnecessary conversions, as
illustrated in Table 3-4.

Table 3-4. RDMS Data Types and Equivalent Language Definitions

RDMS 1100 UCS COBOL ASCII COBOL ASCII FORTRAN

Column Definition Data Data or UCS
Definition Definition FORTRAN

CHARACTER(n) X(n) X(n) CHARACTER*n

A(n) A(n)
NCHARACTER(n) X(n) usage (no equivalent) NCHARACTER*n
disp-2
DECIMAL(d,$) S9(d-s-1)v9(s) S9(d-s-1)v9(s) INTEGER
sign leading sign leading REAL
separate separate DOUBLE
PRECISION
NUMERIC(d,$) S9(d-s)v9(s) S9(d-s)v9(s) INTEGER
usage binary usage comp REAL
DOUBLE
PRECISION
INTEGER S9(10) usage S9(10) usage INTEGER
binary comp
S1(36) usage S1(36)
binary-1
SMALLINT S9(5) usage S9(5) usage INTEGER
binary comp
S1(18) usage S1(18)
binary-1
REAL usage comp-1 usage comp-1 REAL
DOUBLE usage comp-2 usage comp-2 DOUBLE
PRECISION PRECISION
FLOAT(p) usage comp-1 usage comp-1 REAL
usage comp-2 usage comp-2 DOUBLE
PRECISION

3-18 SG 3605
 Using SQL in Application Programs

RDMS 1100/2200 Terminology

Before you can write a program that incorporates SQL commands in the 1100/2200
environment, there are a few more terms that must be defined. They relate your program
to its environment.

Cursor
The definition of the data to retrieve. Used only to return results to a program.

Currency
Currency indicates which row of the cursor is being accessed. This pointer is needed as a
way to keep track of the rows being returned to a COBOL or FORTRAN program, one at
a time.

Application Group
A software partition that acts as an independent database system and has its own system
and data files and its own recovery environment. Applications groups do not share data
with other application groups and there is no coordinated locking or recovery between
application groups. As a programmer, you need to know the name or alias of the
application group to which your program belongs.

Thread
A thread is a work session established with UDS. It defines your UDS environment.
Each thread exists in one and only one application group.

Program Flow

Figure 3-2 illustrates the typical program flow of control and some of the SQL commands
that are involved.

SG 3605 3-19
Using SQL in Application Programs

Start

BEGIN THREAD

DECLARE CURSOR
SELECT

Thread OPEN CURSOR

Under
UDS
Control FETCH

Loop
Process Record

END THREAD

End Program

Figure 3-2. Program Flow with SQL Commands

3-20 SG 3605
 Using SQL in Application Programs

BEGIN THREAD Command

BEGIN THREAD [thread-name]

FOR [APPLICATION] application-name [options]

• Unisys extension to register thread with UDS 1100

- Sets up working storage and environment for subsequent SQL commands

- Specifies application group or alias name for application group to which thread
belongs

- Specifies what type of recovery to use during updates

• Retrieval options

— READ/RETRIEVE

Database access is read only

— Attempt to update data generates an error

• Message options

— UDSMSG or UDSMESSAGE or UDSMESSAGES

— Messages from UDS 1100 are returned immediately

Examples:

BEGIN THREAD mythread FOR app1

BEGIN THREAD FOR UDSSRC READ

BEGIN THREAD t1 FOR UDSSRC READ UDSMSG

BEGIN THREAD FOR app2 RETRIEVE UDSMESSAGE

SG 3605 3-21
Using SQL in Application Programs

DECLARE CURSOR Command

DECLARE cursor-name CURSOR

[FOR] select-clause
[ORDER BY sort-specification]

• Associates a named data area with a particular query

- Defines a conceptual table containing the rows returned before they are passed to
the program, one row at a time

- A single program may have multiple declared cursors

- A cursor must be declared before any other references can be made to it

• SELECT clause specifies which data to return

• ORDER BY clause controls how result is displayed

Examples

DECLARE empl CURSOR

SELECT * FROM emp

DECLARE ci CURSOR
SELECT dname, dno FROM dept
ORDER BY dno

DECLARE sal_cursor CURSOR

SELECT ename, sal FROM emp
WHERE dno = 200
ORDER BY 2 DESC

3-22 SG 3605
 Using SQL in Application Programs

OPEN CURSOR Command

OPEN cursor-name

• Opening the cursor evaluates the query in the DECLARE CURSOR command

- Sets the cursor's currency pointer to just prior to the first row

- Sorts the rows if DECLARE CURSOR had ORDER BY clause

• Performed implicitly by first FETCH for ISQL

• Required for static ESQL

Example:

OPEN SAL CURSOR

SG 3605 3-23
Using SQL in Application Programs

FETCH Command

FETCH cursor-name INTO variable-specification-list

FETCH retrieves the next row of the cursor relative to the currency pointer

— First FETCH retrieves the first row returned based on the cursor definition

- Usually used in a program loop to read all the rows

• Where variable-specification-list is

A list of placeholder variables, one for each column in the cursor (ISQL)

— A set of embedded variables, one for each column in the cursor (ESQL)

ISQL Example:

ENTER MASM 'ACOB$RDMR' USING

'FETCH dept_cursor INTO $p1, $p2, $p3 ;',
err-stat, aux-info, rr-dno, rr-dname, rr-loc

ESQL Example:

EXEC SQL
FETCH dept_cursor INTO :rr-dno, :rr-dname, :rr-loc
END-EXEC.

3-24 SG 3605
 Using SQL in Application Programs

Passing Information Between a Program and RDMS

Figure 3-3 illustrates how data is passed between a program and RDMS.

Note: Only interpretive SQL requires the placeholder variables, $P1, $P2,and $P3.

Host Program Variables

Database applications that use third-generation programming languages will need
program variables declared in the host program. Host program variables are used to pass
information between your program and RDMS.

The host program variable's size and data type is determined from the RDMS column it
corresponds to. You can find this information in the table definition. The size and data
type of your host program variables must match the size and data type of the table
declarations .

You need a host program variable for each column you are inserting values into,
retrieving values from, or comparing values against.

Placeholder Variables in ISQL

ISQL statements cannot directly reference program variables, so placeholders are the
means of passing values between RDMS 1100 and your program. Placeholder variables
are named according to their position in the SQL string in which they are passed; three
placeholder variables would be named $P1, $P2, and $P3. Your call to RDMS must have a
host, program variable for each placeholder variable passed as an additional parameters.

Embedded Variables in ESQL

ESQL allows direct references to host program variables, known as embedded variables.
According to the SQL standard, you must prefix the host program variable name with a
colon (:) and embed it in the command string.

Parameter Passing Limitations

The ASCII COBOL (ACOB) compiler limits the number of parameters passed on an
ENTER MASM call to 31.

The UCS COBOL (UCOB) compiler limits the number of parameters passed to 250.

SG 3605 3-25
Using SQL in Application Programs

ISQL - Placeholder and Host Program Variables

DECLARE dept_cursor CURSOR

SELECT dno, dname, loc
FROM dept

DNO DNAME LOC

Decimal(4) Character(12) Character(10)
100 ACCOUNTING ATLANTA
Table 200 SALES NEW YORK
Data 300 MARKETING DETROIT
400 DISTRIBUTION PRINCETON

ISQL
Requires
Placeholder $P1 $P2 $P3
Variables

Host Program rr-dno rr-dname

Variables

WORKING STORAGE.
01 record-retrieved.
05 rr-dno PIC 9(4).
05 rr-dname PIC X(12).
05 rr-loc PIC X(10).

ENTER MASM 'ACOB$RDMR' USING

'FETCH dept_cursor INTO $P1, $P2, $P3
err-stat,
aux-info,
rr-dno,
rr-dname,
rr-loc

Figure 3-3. Cursor-Program Variable Interaction

3-26 SG 3605
 Using SQL in Application Programs

END THREAD Command

END THREAD

• Disconnects thread from UDS Control

• A Unisys extension command needed to end a work session

• Working storage and environment discarded

• Commits database changes if in update mode

SG 3605 3-27
Using SQL in Application Programs

USE DEFAULT QUALIFIER Command

USE DEFAULT QUALIFIER qualifier-name

• Unisys SQL extension

• Establishes a new default table qualifier

- Used for duration of thread or until another default is specified

- System default qualifier is RDMS for all tables referenced

• Table specification

- Format is: [qualifier.]table-name[:version-name]

- Table acquires qualifier when created

• Useful when referencing

- Multiple tables with same qualifier

- Table-names with long qualifiers

Examples:

USE DEFAULT QUALIFIER VJP A subsequent reference to table EMP references

VJP.EMP

EMP Refers implicitly to RDMS.EMP

VJP.EMP Refers explicitly to VJP.EMP

3-28 SG 3605
 Using SQL in Application Programs

Exercise 3-1

1. Identify which variable names are valid and which are invalid in RDMS.

MY-TABLE

EMPCURSOR

Table*2

Another Cursor

2. If the following columns were defined in a table called CAR SALES, how would the
program variables into which they were to be retrieved be declared?

REG_NUMBER DECIMAL(8)

MAKE_AND_MODEL CHARACTER(10)

DATE_PURCH DECIMAL(7)

SALE_PRICE DECIMAL(6)

REP_CODE SMALLINT

COMMISSION NUMERIC(7,2)

3. Put a number before each command below to indicate the order in which they would
appear in an application program.

OPEN CURSOR

END THREAD

DECLARE CURSOR

BEGIN THREAD

FETCH

SG 3605 3-29
Using SQL in Application Programs

4. Code the complete ESQL or ISQL command to define a cursor that selects the
employee number, employee name, and total compensation from the EMP table.

5. Code the ESQL or ISQL call to fetch the rows retrieved in item 4 above. What
assumptions must you make about the WORKING-STORAGE section of your program?

3-30 SG 3605
 Using SQL in Application Programs

Lab 3-1

Your instructor will discuss your particular lab set-up and requirements with you.
Appendix A is provided to keep track of necessary system usage information.

Using the ACOB or UCOB skeleton provided, write a program to display the employee
number, name, position, and hire date of all employees in Department 200. You must
complete the working storage section and the appropriate ISQL or ESQL commands.

A hard copy of the ACOB skeleton for the interpretive SQL interface (ISQL) is in
Appendix C of this student guide. The ACOB program example in Appendix D should help
you decide what work needs to be done.

Similarly, a hard copy of the UCOB skeleton for working with embedded SQL (ESQL) is
in Appendix E, and is followed by a complete program in appendix F.

In lab you will edit the on-line skeleton program, compile, and execute the program. The
on-line program skeleton names are ACOBSKEL and UCOBSKEL. Information on
compiling, collecting or linking, and executing your program can be found in Appendix A.

The flow chart in Figure 3-4 illustrates the basic steps for the examples in the appendixes
which read and display all of the records in the employee table EMP. Your program logic
will be very similar.

SG 3605 3-31
Using SQL in Application Programs

Start

BEGIN
THREAD

DECLARE
CURSOR

OPEN
CURSOR

END
THREAD

Figure 3-4. Flow Chart of an SQL Retrieval Program

3-32 SG 3605
 Using SQL in Application Programs

Error Handling Choices

• RDMS error handling

- Each SQL call returns an error status to a 4 character field

- Normal completion = 0000

— End-of-cursor = 6001

- Any other value is error

• Automatic error printing

BEGIN THREAD option of UDSMSGTUDSHESSAGE

- Immediately prints an error message to output device

- No other code necessary

- Prints messages from UDS, UREP, and RDMS

• Programmer-controlled error handling

- Check error status after each SQL command

— If error, use GETERROR command to retrieve error text

- Retrieves messages detected by RDMS, UREP, and UDS Control

- Print messages when and where you want

• ESQL provides new error handling features

- Variable SQLCODE for status information

- BEGIN DECLARE command for ESQL variables

- Global WHENEVER command for execution errors

SG 3605 3-33
Using SQL in Application Programs

RDMS Error Handling

Every RDMS call, except an SQL GETERROR call, returns an error status code and
auxiliary information about the completion of the call. The error-status item is a 4 digit
variable in the data declarations of the host program. The auxiliary-information item in
COBOL is declared as a multipurpose, one-word, numeric item.

You can test the value of error-status variable then decide how to handle the error. A
status equal to 6001 indicates a no-find or an end-of-data condition. After fetching from a
cursor, this is not an error. Continue to the next processing step.

The auxiliary information variable returns additional information about the error status.
When the error status is not between 6000 and 6999, it returns the column number in the
command where the syntax error was located. When the error status is between 6000
and 6999, auxiliary information contains other command status and rollback information.

ISQL COBOL Example:

MOVE 'Command string ;' TO ROOM

ENTER MASM 'ACOB$RDMR' USING ROOM, ERR-STAT, AUX-INFO
IF ERR-STAT NOT = 0000 PERFORM ERROR-PRINT

ERROR-PRINT
DISPLAY 'RDML COMMAND = 'ROOM
DISPLAY 'ERROR STATUS = ERR-STAT
DISPLAY 'ERROR COLUMN = 'AUX-INFO

FORTRAN Example:

rcom = command-string
CALL F$RDMR (rcom, erstat, auxinf)
IF (erstat .NE. '0000') THEN
PRINT '(1X,A)', 'RDMS command: ',rcom
PRINT '(A26,I3)', 'Error Found at Column ', auxinf
PRINT '(A16,A4)', 'Error Status: ', erstat
ENDIF

3-34 SG 3605
 Using SQL in Application Programs

GETERROR Command

GETERROR INTO variable-list

• Retrieves the error message from an SQL command

• Unisys extension

• Details on variable-list

- Provides a list of variables separated by commas

- Each variable must correspond to a host program variable

- Each host program variable must hold at least 132 ASCII characters

Five host program variables usually retrieve entire message

• Retrieved message must be printed or moved before executing the next SQL command

ESQL Example:

GETERROR INTO :err(1), :err(2), :err(3), :err(4), :err(5).

SG 3605 3-35
Using SQL in Application Programs

Example Using GETERROR (ACOB)

WORKING-STORAGE SECTION.

01 err-stat PIC 9999.

01 aux-info PIC S9(18).
01 error-text.
05 err-s PIC X(132) OCCURS 5 TIMES.
01 i PIC 9999.
01 err-end-flag PIC 9999.

PROCEDURE DIVISION.

MOVE 'BEGIN THREAD T1 FOR udssrc READ ;' TO rcom.

ENTER MASM 'acob$rdmr' USING rcom, err-stat, aux-info.
IF err-stat NOT = 0 GO TO error-para.

error-para.
MOVE 0 TO err-end-flag.
PERFORM do-geterror UNTIL err-end-flag NOT = 0.
ENTER MASM 'acob$rdmr' USING 'END THREAD ;',
error-stat, aux-info.
DISPLAY "Abnormal end of program." UPON PRINTER.
STOP RUN.

do-geterror.
MOVE 'GETERROR INTO $P1, $P2, $P3, $P4, $P5;' TO rcom.
ENTER MASM 'ACOB$RDMR' USING rcom, err-stat,
aux-info, err-s(1), err-s(2), err-s(3), err-s(4), err-s(5).
PERFORM print-error VARYING i FROM 1 BY 1 UNTIL i > 5 .

print-error.
If err-s(i) NOT = SPACE
DISPLAY err-s(i) UPON PRINTER
ELSE
MOVE i
MOVE 1 TO err-end-flag.

3-36 SG 3605
 Using SQL in Application Programs

ESQL Error Handling

If your ESQL error handling must conform to the current standard, you must define the
variable SQLCODE, use the global WHENEVER command, and define a new data section
with the BEGIN DECLARE command.

The variable SQLCODE must be declared by that name in the Data Division of your
program. It holds the completion status of each executed SQL command. A value of 0
indicates normal completion, +100 means no data or no more data, and a negative value
indicates an error has occurred.

You will use the WHENEVER command to transfer control to a particular paragraph
should a no-find or error condition occur when executing an SQL command.

A new section must define variables you use in ESQL statements. This section, also a part
of the SQL standard, is written in the Working-Storage Section as an embedded
command. It is defined with the Begin Declare command.

A UCOB compiler option, LEVEL/SQL, can be used on the compiler call to flag
nonstandard commands. You can enter this keyword option as you would any other, for
example:

@UCOB,ES SQL.LAB3-1/UCOBSKEL,TPKUCOB„,NARROW,LEVEUSQL

If portability is not a concern, the variables ERROR-STATUS and AUX-INFO in the

structure RDMCA group item can be used to obtain the command status and auxiliary
information returned by RDMS. You must code the variable names and structure name in
a precise way in the Declare Section in the Working-Storage Section. You will use the two
variables along with the GETERROR command just as you would have in ISQL, but you
do not pass them to RDMS with each SQL command.

SG 3605 3-37
Using SQL in Application Programs

Data and Commands for ESQL Error Handling

• SQLCODE contains completion status for each executed SQL command

• WHENEVER command

- Branch on detected SQL execution error

- Branch on no-find condition

• BEGIN DECLARE command for variables used in SQL commands

• LEVEL/SQL flags nonstandard SQL

• RDMCA group item

— Contains variables ERROR-STATUS and AUX-INFO

— Placed in Declare Section

3-38 SG 3605
 Using SQL in Application Programs

WHENEVER Command

WHENEVER { SQLERROR I NOT FOUND }.

{ CONTINUE I { GOTO I GO TO } paragraph-name }

• Detected conditions

— SQLERROR means SQLCODE is negative, an abnormal error status is returned by

RDMS

— NOT FOUND means SQLCODE = +100, a "not found" status returned by RDMS

• Action to take

- CONTINUE means to execute the next program statement

- GO TO/GOTO paragraph-name specifies a paragraph to branch to for the detected

condition

• A WHENEVER command affects all ESQL commands that follow it, until another
WHENEVER command for the detected condition

• To avoid an infinite loop in the error handling paragraph, execute a WHENEVER

SQLERROR, CONTINUE

Examples:

EXEC SQL WHENEVER SQLERROR CONTINUE END-EXEC.

EXEC SQL
WHENEVER SQLERROR GO TO rdms-error-para
END-EXEC.

SG 3605 3-39
Using SQL in Application Programs

ESQL Data Area Declarations

Variable SQLCODE

• Returns the status of each executed embedded SQL command

0 = Successful completion
+100 No data, no more data
- Negative = error

• Does not return a status for GETERROR

BEGIN DECLARE Command

EXEC SQL BEGIN DECLARE SECTION END-EXEC

• Marks a data area for variables used in SQL commands in the COBOL Working-
Storage Section

• Required by SQL standard, not by RDMS

RDMCA with Variables ERROR-STATUS and AUX-INFO

• RDMS uses variables automatically to return status information for each SQL
command, except GETERROR

• Must be defined in a prescribed way

3-40 SG 3605
 Using SQL in Application Programs

Example Using GETERROR (UCOB)

WORKING-STORAGE SECTION.

•
EXEC SQL BEGIN DECLARE SECTION END-EXEC.
01 SQLCODE PIC S9(9) USAGE BINARY.,
01 RDMCA.
05 ERROR-STATUS PIC 9(4). CO,
05 AUX-INFO PIC S9(9) USAGE BINARY.,
EXEC-SQL END DECLARE SECTION END-EXEC.

01 error-text.
05 err-s PIC X(132) OCCURS 5 TIMES.
01 err-index PIC 9.

PROCEDURE DIVISION.

EXEC SQL
WHENEVER SQLERROR GO TO rdms-error-para
END-EXEC.

rdms-error-para.
EXEC SQL WHENEVER SQLERROR CONTINUE END-EXEC.
DISPLAY '***SQLCODE = ' SQLCODE UPON PRINTER.
DISPLAY '***AUXILIARY INFO = ' AUX-INFO UPON PRINTER.
DISPLAY '***ERROR STATUS = ' ERROR-STATUS UPON PRINTER.
GETERROR INTO :err-s(1), :err-s(2), :err-s(3),
:err-s(4), :err-s(5).
PERFORM VARYING err-index FROM 1 BY 1 UNTIL err-index = 6
DISPLAY err-s(err-index) UPON PRINTER
END-PERFORM.
PERFORM end-thread-to-uds.
PERFORM termination-para.

SG 3605 3-41
Using SQL in Application Programs

Lab 3-2

Edit, recompile, and rerun the program from Lab 3-1:

a. With and without UDSMSG on.

b. With and without GETERROR as part of the ERROR-PRINT paragraph.

Insert errors, such as too few program variables or inconsistent data types for program
variables to receive query results. Then execute the program to see the types of error
messages returned.

Of course, if you are working in ESQL most of your errors are detected and flagged at
compilation time. You should force several compile time errors in order to become familiar
with the error messages.

3-42 SG 3605
4
More WHERE Clause
Operators
 More WHERE Clause Operators

Module 4
More WHERE Clause Operators

Module Objectives 4-3

Benefit Statement 4-3

Materials 4-3

Testing Conditions 4-4

Multiple Conditions 44

NULL Values 4-5

Testing NULL Values 4-6

Logical Operators 4-7

Truth Tables for Three-Valued Logic 4-8

AND 4-8
OR 4-8
NOT 4-8

BETWEEN Operator 4-9

Order of Precedence of Operators 4-10

LIKE Operator 4-11

LIKE Examples 4-12

Exercise 4-1 4-13

Lab 4-1 4-15

SG 3605 4-1
 More WHERE Clause Operators

Module Objectives

Upon completion of this module, you should be able to

• Formulate queries that fully exploit the WHERE clause.

- Using multiple conditions.

- Using the logical operators AND, OR, and NOT.

Using NULL values.

- Using the BETWEEN...AND, LIKE, and IS NULL operators.

Benefit Statement

In Module 2 you learned the elements of simple queries. This module builds upon that
module with extensions to the WHERE clause. You will now be able to construct WHERE
clauses that contain multiple conditions, test for NULL values, and use pattern matching.

Materials

• OS 1100 UDS Relational Data Management System (UDS RDMS 1100) and IPF SQL
Interface End Use Guide

• OS 1100 UDS Relational Data Management System (UDS RDMS 1100) SQL
Programming Reference Manual

SG 3605 4-3
More WHERE Clause Operators

Testing Conditions

A condition is an expression that evaluates to TRUE, FALSE, or NULL (unknown). SQL

queries return those rows of a table for which the data make the conditions in the
WHERE clause evaluate to TRUE.

Multiple Conditions
You have already written simple queries using a single condition in Module 2. But in real
life, you must be able to express more complex sets of conditions. In SQL you can
construct a query that is as complex as you require by combining conditions with logical
operators.

• You can create conditions using the following operators:

— AND

- OR

NOT

— BETWEEN ... AND

- IS [NOT] NULL

- [NOT] LIKE

• You can control the appearance of the output using ORDER BY.

4-4 SG 3605
 More WHERE Clause Operators

NULL Values

• Use when actual value for a column is unknown

• Use NULL instead of inserting dummy values

• NULL values are not equivalent to zero or spaces

• Expressions with NULL will evaluate to NULL

10 * NULL is NULL

• There are special ways to handle displaying NULL values using indicator variables

SG 3605 4-5
More WHERE Clause Operators

Testing NULL Values

IS [ NOT ] NULL

• Tests whether column data value is known

• Allows different handling of result when no value is present

— Print "Not applicable" instead of NULL in report

• Any arithmetic comparison with NULL is unknown

Example 1. List all employees who do not have a manager.

SELECT * FROM EMP

WHERE MGR IS NULL

Example 2. List all employees who are eligible for commissions.

SELECT * FROM EMP

WHERE COMM IS NOT NULL

Example 3. Why is the above query different from the following?

SELECT * FROM EMP

WHERE NOT COMM = 0

4-6 SG 3605
 More WHERE Clause Operators

Logical Operators

• AND - Both conditions are true

WHERE SAL > 1200

AND DNO = 200

• OR - Either or both conditions are true.

WHERE SAL > 1200

OR DNO = 200

• NOT - Reverses the truth value of an expression

- NOT must precede the entire condition

- It is not used to negate the operator

- Valid

WHERE NOT JOB = 'CLERK'

- Invalid

WHERE JOB NOT = 'CLERK'

• The word WHERE appears only once, no matter how many conditions are used

SG 3605 4-7
More WHERE Clause Operators

Truth Tables for Three-Valued Logic

AND
Table 4-1. AND Truth Table

AND TRUE FALSE NULL

TRUE TRUE FALSE NULL
FALSE FALSE FALSE FALSE
NULL NULL FALSE NULL

OR
Table 4-2. OR Truth Table

OR TRUE FALSE NULL

TRUE TRUE TRUE TRUE
FALSE TRUE FALSE NULL
NULL TRUE NULL NULL

NOT
Table 4-3. NOT Effect

NOT
TRUE FALSE
FALSE TRUE
NULL NULL

4-8 SG 3605
 More WHERE Clause Operators

BETWEEN Operator

WHERE x BETWEEN a AND b

• Is a comparison operator, not a logical operator

• Logical equivalent of

x >. a AND x <= b

Example:

SELECT * FROM EMP

WHERE HIREDATE BETWEEN 850101 AND 900101

SG 3605 4-9
More WHERE Clause Operators

Order of Precedence of Operators

Table 4-4. Operators and Order of Precedence

Operator Symbol Order

Arithmetic Unary + - 1
*1 2
+- 3
Comparison <> (not equal) 4
< 4
> 4
<= 4
>= 4
= 4
IS [ NOT ] NULL 4
[ NOT ] LIKE 4

Boolean NOT 5
AND 6
OR 7

Table 4-4 shows the operators RDMS accepts and the order of precedence each has been
assigned. You can use parentheses to change the order of precedence as needed.

4-10 SG 3605
 More WHERE Clause Operators

LIKE Operator

column-spec [ NOT ] LIKE { string-1 I variable-spec-1 }

[ { ESCAPE string-2 I variable-spec-2 } ]

column-spec A character or ncharacter column containing the data

string-1 A pattern enclosed in apostrophes

string-2

variable-spec-1 A data variable, either an embedded variable or placeholder

variable-spec-2 variable, that contains a pattern

• Performs character string pattern matching in a WHERE clause

— Case-sensitivity is significant

If column is NULL it evaluates as false

• Exact match search should use "=" operator

• Partial string search should use LIKE

Table 4-5. Wild Cards

Wildcard Character Matches

_ (underscore) Any single character
% (percent) Any combination of 0
or more characters

• Two characters, "%" and "_ , behave as "wild cards" for the LIKE operator

• Search string may not contain consecutive "%"

• Search for strings that include "%" or "_" by preceding them with escape character
specified

SG 3605 4-11
More WHERE Clause Operators

LIKE Examples

Example 1. List the employee information from the EMP table for employees whose
names have five characters..

SELECT * FROM EMP

WHERE ENAME LIKE '

Example 2. List the employee information from the EMP table for all employees whose
last names begin with "W'.

SELECT * FROM EMP

WHERE ENAME LIKE 'W%'

Example 3. List employees from EMP whose names have an "A" as the third from last
character.

SELECT * FROM EMP

WHERE ENAME LIKE '%A'

Example 4. List employee information from the EMP table for employees whose names
have an "A" followed by an "S".

SELECT * FROM EMP

WHERE ENAME LIKE "Y0A%S%

Example 5. From a real estate system, search for a house with a "9% assumable" in
the description column.

SELECT * FROM HOUSES

WHERE DESC LIKE '9#% assumable' ESCAPE '#'

4-12 SG 3605
 More WHERE Clause Operators

Exercise 4-1

Indicate whether the following queries have valid or invalid syntax. If invalid, correct the
query so that it has valid syntax.

1. SELECT * FROM EMP

WHERE MGR = NULL

2. SELECT * FROM DEPT

WHERE DNO = 200 OR
WHERE DNO = 400

3. SELECT * FROM EMP

WHERE NOT ENAME LIKE 'F%' AND
HIREDATE BETWEEN 850101 AND 870631

4. SELECT * FROM EMP

WHERE EMPNO > 5200 AND
JOB NOT = 'CLERK'

5. SELECT * FROM DEPT

WHERE LOC = 'ATLANTA' OR 'DETROIT'

SG 3605 4-13
More WHERE Clause Operators

Write queries to display all the employee information for employees who satisfy the
following conditions:

6. All clerks in either department 200 or 400.

r

373

7. All employees hired after Jan 1, 1986 whose total compensation per month is more
than $2500.

8. All sales reps and accountants who earn less than $2300 and more than $2000.

( (- 1'/-••

9. All employees whose name have four or fewer letters.

f;- .3,- • •

ttk
10. All employees not in department 100 holding any job but president or clerk.
/„..c

4-14 SG 3605
 More WHERE Clause Operators

Lab 4-1

For this lab you will begin with the program as you completed it from Lab 3-1.
You will modify it so that it now uses a more complex query. The two queries that you
wrote as answers to questions 7 or 8 in Exercise 4-1 are good candidates for the requested
"complex query" for this exercise.

SG 3605 4-15
5
The IPF SQL Interface
 The IPF SQL Interface

Module 5
The IPF SQL Interface

Module Objectives 5-3

Benefit Statement 5-3

Materials 5-3

The IPF SQL Interface to RDMS 1100 5-4

Using IPF SQL 5-5

Invoking IPF 5-5

Sample IPF SQL Session 5-7

Creating an SQL File 5-8

Sample IPF SQL File 5-9

Executing an SQL File 5-9

Saving Query Results in a File 5-10

Lab 5-1 5-11

SG 3605 5-1
 The IPF SQL Interface

Module Objectives

Upon completion of this module, you should be able to

• Invoke and use the IPF SQL interface to RDMS for simple queries.

- Toggle between line and fullscreen mode.

- Interactively submit SQL commands.

- Create, save and run IPF SQL files.

Benefit Statement

The IPF SQL interface is easy to learn and use; it will greatly enhance your mastery of
the SQL syntax. You will use this interface for ad hoc queries as well as for prototyping
applications being developed.

Materials

OS 1100 UDS Relational Data Management System (UDS RDMS 1100) and IPF SQL
Interface End Use Guide

SG 3605 5-3
The IPF SQL Interface

The IPF SQL Interface to RDMS 1100

• IPF SQL provides online access to RDMS data

• Interactive query and update of tables

• Interfaces with RDMS to parse and execute SQL

• Has online help facility to provide information on commands

— Type "SQL?" to obtain general information on SQL

Type "SQL command ?" to obtain command information, where the command is
represented by the first one or two keywords

— Type "?" to obtain more information until you receive the end of message notice

5-4 SG 3605
 The IPF SQL Interface

Using IPF SQL

SQL commands can be used with IPF commands in an .IPF session. You can:

• Intersperse SQL and IPF commands when working interactively at a terminal.

• Store SQL commands in a file and execute them all at once

- The file can be saved, edited, and re-executed at any time

- Only SQL commands may be included in the file

- All commands must be preceded by SQL to tell IPF they are SQL commands

• Continue a command to the next line by using the IPF continuation character.

- It is the ampersand (&) unless you change it.

- System variable $CONTCHAR sets the continuation character.

• Omit the SQL command termination character (;) in IPF

Invoking IPF

• Invoked by typing "@IPF" after the SOE.

• In IPF you have a choice of two screen modes:

- Line mode

- Full-screen mode

• Any IPF or SQL command may be entered following the >-C-> prompt.

• Figure 5-1 illustrates the sequence of commands executed in a typical IPF SQL
session.

SG 3605 5-5
The IPF SQL Interface

Invoke IPF

Set Datamanager

BEGIN THREAD

Other SQL
Commands

END THREAD

Exit IPF

Figure 5.1. Flow of an IPF SQL Session

5-6 SG 3605
 The IPF SQL Interface

Sample IPF SQL Session

>@IPF TPF$.
>IPF 1100 5R1 02/19/93 09:26:11
>-C->$DATAMANAGER := RDMS
>-C->SQL BEGIN THREAD Ti FOR UDSSRC READ
>-C-->SQL USE DEFAULT QUALIFIER ROSVSCHEMA1
>The USE command completed successfully.
>-C->SQL SELECT * FROM DEPT WHERE DNO >= 200
>-DNO- ---dname---- ---loc----
> 200 SALES NEW YORK
> 300 MARKETING DETROIT
> 400 DISTRIBUTION PRINCETON
>3 records were selected.
>-C->SQL SELECT ENAME, HIREDATE, SAL * 12 FROM EMP &
>-C--&>WHERE HIREDATE BETWEEN 890101 AND 900101
>--ename--- hiredate
>TURNER 890515 10800
>JOHNSON 891221 11700
>FORD 890316 28500
>3 records were selected.
>-C->SQL END THREAD
>-C->logoff
>END IPF

• Start of entry (SOE) character is represented by the greater than symbol (>) above

• $DATAMANAGER can be abbreviated to $DATA

SG 3605 5-7
The IPF SQL Interface

Creating an SQL File

• What is an SQL file?

- A place to edit, save, and execute SQL commands

Typically a symbolic element containing SQL command only

- It may not contain a mixture of IPF and SQL commands

• How is it created?

With the IPF editor in either line or fullscreen mode

For fullscreen edit mode, type "MODE SCREEN"

Return to line mode by typing "MODE LINE"

Screen mode is recommended for entering and editing text, and line mode for
executing the file or interactive SQL commands

• How do I get started?

- Begin a new file by typing "NEW filename" on the command line at the top

The file is opened and is in input mode

Enter text into the editor

Save the file by entering "SAVE" on the command line

- If the file has already been saved, then you must use the "REPLACE" (REPL)
command to save it

5-8 SG 3605
 The IPF SQL Interface

Sample IPF SQL File

SQL BEGIN THREAD FOR UDSSRC READ

SQL USE DEFAULT QUALIFIER ROSVSCHEMA1
SQL SELECT ENAME, DNO, MGR, SAL * 12 FROM EMP &
WHERE DNO = 200
OR DNO = 400
SQL SELECT * FROM EMP
WHERE HIREDATE > 870101
AND SAL > 2000
SQL END THREAD

Executing an SQL File

To execute a file, type "SQL FTLE filename" on the command line. When you submit a
FILE command, IPF SQL accesses the file or element specified, then reads and executes
the SQL commands. It executes all the commands even if it detects errors as it proceeds.
IPF returns error messages interspersed with the results of the commands it executes.

Files are better executed in line mode since the screen otherwise returns to the previous
edit screen, overlaying the query results. To enter line mode, type "MODE LINE" on the
command line.

SG 3605 5-9
The IPF SQL Interface

Saving Query Results in a File

>-C->CREATE FILE=OUTFILE. ACCESS=PUBLIC

>-C->OUT FILE=OUTFILE.
>-C->SQL SELECT * FROM EMP
>-C-->OUT FILE=TERMINAL.
>-C->OLD FILE=OUTFILE.

• Steps illustrated above

First command creates a data file to receive the output

- Second command directs output to the file

- Third command returns results to the output file

- Fourth command redirects the output the terminal

Fifth command copies the contents of the output file to the IPF workspace for
viewing

• Output file can be directed to a printer for a hard copy

5-10 SG 3605
 The IPF SQL Interface

Lab 5-1

1. Invoke the on-line help facility to obtain information on SQL in general (SQL ?) and
SQL commands (SQL command-name ?) in particular.

2. Enter a SELECT command before a BEGIN THREAD to see the kind of error
messages that are returned.

3. Use interactive IPF SQL to perform the following queries:

a. Display all information about all employees in the EMP table.

b. List all information about the departments in the DEPT table.

c. Display the employee number, name, position, and hire date of all employees in
department 200.

d. Display the name, salary, and employee number of all clerks.

4. Create, save, and execute an IPF SQL file that includes any two queries from
Exercise 4-1.

SG 3605 5-11
6
The MAPPER
Relational Interface
 The MAPPER Relational Interface

Module 6
The MAPPER Relational Interface

Module Objectives 6-3

Benefit Statement 6-3

Materials 6-3

MAPPER Relational Interface (MRI) 6-4

Invoking RDI 6-6

RDI Menu Selections 6-7

Environment Menu Selections 6-8

Selection Example 1 - Menu Access 6-9

Selection Example 2 - Freeform Entry 6-11

Exiting MAPPER 6-13

Embedding Commands in MAPPER Runs 6-14

MAPPER Run Example Using FCH 6-15

Lab 6-1 6-16

Lab 6-2 6-17

SG 3605 6-1
 The MAPPER Relational Interface

Module Objectives

Upon completion of this module, you should be able to

• Invoke the MRI run to retrieve and manipulate data from an RDMS database.

Benefit Statement

An alternative to the IPF interface is the MAPPER Relational Interface (MRI). You will
find this interface even easier to use to obtain ad hoc reports, with its menu-driven
approach. Of course, MRI supports more sophisticated database accesses as well, and you
will be introduced to SQL in MAPPER runs as well.

Materials

• MAPPER Relational Interface (MRI) Relational Database Interface (RDI) Operations

Guide

• MAPPER System Run Design Operations Reference Manual

SG 3605 6-3
The MAPPER Relational Interface

MAPPER Relational Interface (MRI)

Provides connectivity to relational databases on various architectures

Released as a separate product

- Uses same menus, regardless of architectural platform

• Consists of seven run statements

- Allows access to relational data

Generates code in SQL

— Performs full range of relational operations

• Provides menu-based interface (RDI run) for ease of use

Submits SQL statements in an easy-to-use format

Displays table data as a MAPPER result

Provides on-line help at all menu levels

• Supplies RDMS information for further manipulation by MAPPER

- Data reduction

- Graphics

- Display/store of data

6-4 SG 3605
 The MAPPER Relational Interface

MAPPER

Request
Manager

MRI
Remote
Execution

SYBASE

DMSII

DB2 INFORMIX

Oracle

Figure 6-1. MRI Access to Data Managers

SG 3605 6-5
The MAPPER Relational Interface

Invoking RDI

Sign-on to MAPPER

• Type RDI on the control line

— Displays the RDI menu

— Allows you to select the SQL command to perform

Figure 6-2 shows RDI entered on the top line, the control line.

rdi
Unisys MAPPER System
• EDMAP 35R1 EA18 *
Station: 8417 System: 1
User-id: DAA
Cabinet: 8

1Report 214eyHlp 3Runs 4 5 6Tasks ?Remote 8Help 9 lagnOff

Figure 6-2. Invoking RDI

6-6 SG 3605
 The MAPPER Relational Interface

RDI Menu Selections

• Pictured in Figure 6-3 is the RDI Menu

• Selection choices

- Top five correspond to SQL commands

- SQL allows you to submit SQL syntax directly to RDMS

Utilities menu allows you to change the design, restrict table access, delete tables,
or work with views

- Detailed help is last selection

• Function keys appear along bottom

— F6 (DBMS) changes the database manager (Default is RDMS)

F7 (Envmnt) changes the RDI environment

— F8 (Help) is the help key

— F10 exits the RDI run

RDI

Tab to your choice and press MIT

Select ►
Join 0-
Insert ►
Delete 0-
Update II.
Create 0-
SQL ►
Utilities menu 10-
Detailed help ►
Rev. Level 3R3

1 2 3 4 5 6DBMS Unumnt BHelp 16Qui t

Figure 6-3. RDI Menu

SG 3605 6-7
The MAPPER Relational Interface

Environment Menu Selections

Pictured in Figure 6-4 is the Environment Menu

• Selection choices

- Results are stored by default to cabinet drawer OA

- Tracking of SQL syntax to a report can be disabled

- ROSVSCHEMA1 is the default qualifier entered for future table references

- Default version, PRODUCTION, has been cleared

• F4 (Return) takes you back to the RDI menu

RDI
Environment menu_
Enter cabinet number for results -> 8 Drawer -> a
Generate SQL syntax on updates? -> y
New default qualifier -> ROSUSCHENA1
New default version -)
Place cursor here and press XNIT ->

1 2 3 4Return 5 6 7 BHelp 9 18Quit

Figure 6-4. Environment Menu

6-8 SG 3605
 The MAPPER Relational Interface

Selection Example 1 - Menu Access

On Figure 6-5 the cursor was positioned to Select and transmitted.

RDI
RDI menu
Tab to your choice and press XMIT
Select ►
Join ►
Insert 00
Delete 00
Update P
Create ►
SQL ►
Utilities menu P
Detailed help ►
Rev. Level 3R3

1 2 3 4 5 6DBMS ?Envmnt SHelp 9 10Quit

Figure 6-5. Selecting SELECT

On the Select Menu pictured in Figure 6-6, the table name EMP was entered and all
columns selected with the *. The cursor was transmitted from the XMIT position.

RDI
Select menu
Retrieve rows from a single relational table.
Table name ->
Column names -> *
Where clause -)
Retrieve all rows? -> y If no, how many? -> 1
Number of rows to skip -> 8 Null Symbol -) *
Extra options? -> n
Place cursor here and press XMIT ->
Press XMIT from here for detailed help ->

1 2 3 4RDIMnu 5 6Specl ? SHelp 9 10Quit

Figure 6-6. Entering the Selection Criteria

SG 3605 6-9
The MAPPER Relational Interface

The results are shown in Figure 6-7 from the generated SQL command.

Line! 5 Ro1110. - RESULT

.DATE 18:26:13 RID 19 FEB 93 DAA
* . .HIREDAT.
*EMPNO. . ENAME :JOB :MGR .DNO .E .SAL :COMM
34.7. • • •
5010 FOSTER SALESREP 5234 200 860714 2700 400
5146 BROWN CLERK 5234 200 871011 1000
5234 WOODWORTH MANAGER 5784 200 790219 3250
5237 ROCKWELL ACCOUNTANT 5743 100 840618 2175
5437 MARTIN SALESREP 5234 200 870326 2500 300
5469 ADAMS SALESREP 5234 200 811102 2250 500
5630 GLASS ACCOUNTANT 5743 100 871220 1850
5702 TURNER CLERK 5743 100 890515 900
5743 LAWSON MANAGER 5784 100 830510 3400
5765 JOHNSON CLERK 5984 480 891221 975
5784 WILLIAMS PRESIDENT 300 781015 5200
5896 SMITH CLERK 5984 400 880913 1058
5942 FORD SALESREP 5234 208 890316 2375 0
5984 HYDE MANAGER 5784 400 850901 2980
END REPORT

1SelMnu 2 3ShwSQL 4Return 5 6Tasks ?View 8 9 10Quit

Figure 6-7. Viewing the Results from the Select

Pressing F3 (ShwSQL) on the previous menu produces the SQL shown in Figure 6-8.
Pressing F3 (SveSQL) from this menu allows you to save the generated SQL in a report
for future use. This was not done in this example.

4 - RESULT
.DATE 19 FEB 93 18:26:12 REPORT GENERATION DAA
- RDI Select, SQL request
SELECT
FROM EMP
END REPORT

1Se 1Mnu 2 3SueSQL 4Return 5 6 7 8 9 lOQui t

Figure 6-8. Using F3 to Show SQL

6-10 SG 3605
 The MAPPER Relational Interface

Selection Example 2 - Freeform Entry

In order to enter your SQL syntax directly, first position to SQL on the RDI menu as
pictured in Figure 6-9 and transmit.

RDI
RDI menu
Tab to your choice and press XMIT
Select ►
Join
Insert ►
Delete
Update ►
Create ►
SQL ►
Utilities menu
Detailed help
Rev. Level 3R3

1 6DBMS 7Envmnt 8Help 9 18Quit

Figure 6-9. Selecting SQL from RDI Menu

On the SQL Menu, merely position the cursor to the XMIT and transmit.

RDI
- SQL menu .
You nag enter a report number, an SQL request name, or leave
both fields blank for freeform entry.
Report ->
SQL request -)
Process request immediately? -> n
Place cursor here and press XMIT ->
Press XMIT from here for detailed help ->

1 2 3 4RDIMnu 5 6 7 8Help 10Qui t

Figure 6-10. SQL Menu

SG 3605 6-11
The MAPPER Relational Interface

On the free form report, enter your SQL statement as you would normally, ending it with —
a semicolon. Press XMIT to have the command processed and then Fl to continue. Figure
6-11 illustrates a SELECT command to be submitted.

Line0. 6 Ro110- - RESULT

.DATE 19 FEB 93 18:38:51 REPORT GENERATION DAA
. Enter a new SQL statement or make any changes to an existing statement.
. Press XMIT at the end of the statement to accept it. Press the
. Resume key to continue. 0.-
*-
SELECT *
FROM EMP
WHERE DNO = 200 ;

END REPORT

1Resume 2Paint 3 4Return 5 6LineCh 7 8 9 10Quit

Figure 6-11. Selection Statement Entered

Figure 6-12 shows the results of the submitted SELECT.

5 - RESULT
.DATE 18:41:04 RID 19 FEB 93 DAA
* • . .HIREDAT.
*EMPNO. ENAME :JOB .MGR .DNO .E .SAL .COMM
*- .
5010 FOSTER SALESREP 5234 200 860714' 2700. 400
5146 BROWN CLERK 5234 208 871011 1000
5234 WOODWORTH MANAGER 5784 200 790219 3258
5437 MARTIN SALESREP 5234 200 870326 2500 300
5469 ADAMS SALESREP 5234 200 811102 2258 500
5942 FORD SALESREP 5234 280 890316 2375
END REPORT

1SQLMnu 2 3ShwSQL 4Return 5 6Tasks ?View 8 9 10Quit

Figure 6-12. Result from Previous SELECT

6-12 SG 3605
 The MAPPER Relational Interface

Exiting MAPPER

Place an X on the control line to exit MAPPER. Figure 6-13 shows the exit from the RDI
menu screen.

X RDI
RDI menu
Tab to your choice and press XMIT
Select 0-
Join 0-
Insert ►
Delete 0-
Update 0.
Create 0-
SQL ►
Utilities menu
Detailed help
help 0-
Rev. Level 3R3

1 2 3 4 5 6DBMS 7Envmnt 8Help 9 Inuit

Figure 6-13. Exiting MAPPER

SG 3605 6-13
The MAPPER Relational Interface

Embedding Commands in MAPPER Runs

• LGN Log On to Relational Database

Use to initiate a session on the relational database

• SQL Submit SQL

Use to pass SQL syntax to the relational database and retrieve data in
MAPPER variables

• FCH Relational Aggregate Fetch

Use to retrieve columns and rows from a relational table

• TRC Trace Relational Syntax

Use to save the SQL syntax generated by run statements

• RAM Relational Aggregate Modify

Use to create tables, insert rows, or update a table

• DDI Data Definition Information

Use to retrieve a table description from a relational database

• LGF Log Off Relational Database

Use to end a session on a relational database

For more information, refer to the MAPPER System Run Design Operation Reference
Manual.

6-14 SG 3605
 The MAPPER Relational Interlace

MAPPER Run Example Using FCH

Figure 6-14 shows an example of a MAPPER run that will perform a simple fetch.

LINE! 1 FMTOP RD- - SHFTO, HLD CHRSR HLD LNO- UNDO- 11124 11.
.DATE 23 FEB 93 88:54:44 RID II 19 AUG 92 DAA
.RBN81 1881760
@LGN,I,Y,8,AJDMS READ
@FCH,8,A,ILREADyI RDMS \
'C * RoSyScHEMAl.DEPT'
8DSP,-0
@LGF,1RDMS
@GTO END .
@I:DSP,-0
END REPORT

Figure 6-14. Fetch All Rows in DEPT

Entering the name of the run on the command line produces the result shown in Figure
6-15.

linen 1 fmtIP rlOP - shft0 hld chr0 hld 10 OP RESULT IP

.DATE 88:57:16 RID 23 FEB 93 DAA
*DNO .DNANE .LOC
ie .
*D4.8 C12 C18
*P
* NULL
188 ACCOUNTING ATLANTA
288 SALES NEW YORE
389 MARKETING DETROIT
488 DISTRIBUTION PRINCETON
END REPORT

Figure 6-15. Result from Run

SG 3605 6-15
The MAPPER Relational Interface

Lab 6-1

Use the following steps to test the basic RDI functionality.

1. Log on to MAPPER following directions provided by your instructor.

2. Invoke the MRI RDI run by typing RDI at the SOE.

3. Move cursor to SELECT and transmit.

4. Complete the SELECT menu as follows:

Table name -> Yourqualifier.EMP

Column Names -> ?
Move cursor to transmit field and press XMIT.

The help screen showing the list of columns is displayed.

5. Select columns to display by typing a number in the first field, indicating the order in
which they should be displayed.

Press XMIT to send the selected columns then press Fl.

6 Press F4 to return to the SELECT menu.

7. Delete the ? from the column name field. Tab to the WHERE field. Enter a ? and
transmit.

The WHERE assistance screen is displayed.

8. Read the information provided.

Press F3 to re-display the list of column names in a corner of the screen for reference.
Press Fl to continue.

9. Type the following in the WHERE field at the bottom of the screen.

WHERE hiredate between 800101 and 891031.

10. Press XMIT then Fl to execute the query. The result is displayed.

6-16 SG 3605
 The MAPPER Relational Interface

Lab 6-2.

Use the MRI interface to display the EMPNO, ENAME, and MGR columns of EMP where
the manager's empno is greater than the employee's empno.

SG 3605 6-17
7
Grouping Functions
 Grouping Functions

Module 7
Grouping Functions

Module Objectives 7-3

Benefit Statement 7-3

Materials 7-3

Grouping Functions 7-4

AS Clause 7-5

Function Examples 7-6

Grouping Functions Usage 7-7

COUNT Function 7-8

GROUP BY Clause Usage 7-9

Grouping on More Than One Column 7-10

WHERE and GROUP BY Clauses Together 7-11

HAVING Clause 7-12

Full Query Specification 7-13

Exercise 7-1 7-14

SG 3605 7-1
 Grouping Functions

Module Objectives

Upon completion of this module, you should be able to

• Formulate queries that group or summarize data retrieved from a table.

- Identify the summarizing functions that can be embedded in SQL commands.

- Rename column output with the AS clause.

— Use the GROUP BY clause.

— Use the HAVING clause.

Benefit Statement

The grouping functions in this module help answer the question, "What is the bottom line
here?" The GROUP BY clause lets you obtain summarized information, and its
companion HAVING clause refines the information you obtain. The AS clause provides
your report with a meaningful name. So what is the bottom line here? More power for
querying your database!

Materials

• OS 1100 UDS Relational Data Management System (UDS RDMS 1100) SQL
Programming Reference Manual

• OS 1100 UDS Relational Data Management System (UDS RDMS 1100) and IPF SQL
Interface End Use Guide

SG 3605 7-3
Grouping Functions

Grouping Functions

Grouping functions operate on a group or set of rows and return one row of summary
information for each group. The default group is the entire table, but other groups may
be specified.

The grouping functions are:

- SUM Computes the total value of the group

- AVG Computes the average value of the group

MIN Computes the minimum value of the group

- MAX Computes the maximun value of the group

- COUNT Counts the number of values in the group

- COUNT(*) Counts the number of rows in the table, including duplicates and
NULLS

Note: COUNT and COUNT(*) are discussed separately in this module.

• Syntax for AVG, MAX, MIN, SUM

SELECT group function( [ ALL] { column-spec I arithmetic-expression } )

FROM table-name

SELECT group function( DISTINCT column-spec)

FROM table-name

Note: An arithmetic-expression must include a column specification.

7-4 SG 3605
 Grouping Functions

AS Clause

select-expression AS title-name

• Use the AS clause to name or rename an item in the select list

Provides a meaningful title for a report

Simplifies sort descriptions in the ORDER BY clause

• Provide a title-name for:

— A column name
- A result of an arithmetic expression
— A constant

A result of a function

Example: Display the names and yearly salaries of all employees in Department 400.

SELECT ENAME, SAL * 12 AS NEWSAL

FROM EMP
WHERE DNO = 400

ENAME NEWSAL

JOHNSON 11700
SMITH 12600
HYDE 34800

SG 3605 7-5
Grouping Functions

Function Examples

Example 1. Compute the total salary for all employees.

SELECT sum(sal) AS "total emp sal"

FROM EMP

total emp sal

32,525

Example 2. Compute the average salary for all employees in the company.

SELECT avg(sal) AS sal_avg

FROM EMP

sal_avg
323.21428571428571429

Example 3. Compute the minimum and maximum salaries paid in the company.

SELECT MIN(SAL) AS minsal, MAX(SAL) AS maxsal

FROM EMP

minsal maxsal
900 5200

Example 4. What is the earliest date on which someone was hired and what was the
most recent date on which someone was hired?

SELECT MIN(HIREDATE), MAX(HIREDATE)

FROM EMP

781015 891221

7-6 SG 3605
 Grouping Functions

Grouping Functions Usage

• There are various names for these functions:

Grouping (or group by) functions

Summarizing functions

Aggregate functions

- Set functions

• Use functions only in a SELECT list or in a HAVING clause

• Functions which apply to a group of records return a single value for the group.

• The DISTINCT keyword used within a grouping function makes the function consider
only distinct values of the expression.

Use the ALL keyword to consider all values including duplicates.

Do not specify DISTINCT or ALL for AVG and SUM functions, since ALL is
assumed.

• Any of the functions may be used with numeric columns

Only MIN, MAX, and COUNT may be used with character columns

You cannot pass a column specification in a variable.

• MIN/MAX functions use the ASCII collating sequence to order and select

• NULL values are ignored in the evaluation of SUM, AVG, MAX and MIN.

SG 3605 7-7
Grouping Functions

COUNT Function

SELECT COUNT { ( * I DISTINCT column-specification )

FROM table-name

• Counts the number of rows returned by a query

• NULL values are ignored

• Asterisk may be used to refer to all columns since all have the same number of rows

Example 1. Display the total number of employees in the EMP table.

SELECT COUNT(*)
FROM EMP

14

Example 2. Count the number of different job titles in the EMP table.

SELECT COUNT(DISTINCT JOB)

FROM EMP

Example 3. Count the number of people who work in department 400.

SELECT COUNT(*)
FROM EMP
WHERE DNO = 400

7-8 SG 3605
 Grouping Functions

GROUP BY Clause Usage

• The GROUP BY clause is used to cluster rows into groups.

A group is a set of records that have the same value in one or more columns.

The column(s) used to define the group must be identified in the GROUP BY
clause

- The grouping function in the SELECT clause then returns one row for each group
instead of one row for the whole table.

For example, grouping by department number causes a row of summary information to be

returned for each department (100, 200, 300, and 400), since those are the distinct values
in the DNO column

• Parameters in the select-list must either be:

- The column(s) used to define the group in the GROUP BY clause, or

- Grouping functions.

— Literals

Valid Example

SELECT COUNT(`), AVG(SAL), DNO

FROM EMP
GROUP BY DNO‘

Invalid Example

SELECT ENAME, AVG(SAL), DNO

FROM EMP
GROUP BY DNO

Note: ENAME is not a grouping function and is not a column-name in the GROUP
BY clause and therefore returns an error.

SG 3605 7-9
Grouping Functions

Grouping on More Than One Column

SELECT group-by-columns or functions

FROM table-name
GROUP BY column1, column2,

• One group is formed for each unique combination of values in the GROUP BY
columns.

• Each row contains summary information for that group.

Example Count the number of people who hold each type of job in each department.

SELECT DNO, JOB, COUNT(*)

FROM EMP
GROUP BY DNO, JOB

--dno-- ---job
300 PRESIDENT 1
200 SALESREP 4
100 CLERK 1
200 CLERK 1
400 CLERK 2
100 MANAGER 1
200 MANAGER 1
400 MANAGER 1
100 ACCOUNTANT 2

7-10 SG 3605
 Grouping Functions

WHERE and GROUP BY Clauses Together

The WHERE clause can be used in conjunction with GROUP BY, to select rows from a
table upon which the grouping functions will be performed.

SELECT group-by-columns or functions

FROM table-name
WHERE conditions
GROUP BY column1, column2,

Example: Determine the average salary for accountants and sales representatives and
the number of each in each department.

SELECT DNO, AVG(SAL), COUNT(*)

FROM EMP
WHERE JOB = 'ACCOUNTANT' OR JOB = 'SALESREP'
GROUP BY DNO

--dno
100 2012.50 2
200 2456.25 4

Note: The average salary in this example was truncated to two decimal places.

SG 3605 7-11
 Grouping Functions

HAVING Clause

SELECT group-by-columns or functions

FROM table-name
GROUP BY column1, column2,
HAVING group-conditions

• Specifies which groups should be returned, after the grouping functions have been
computed

• Similar to a WHERE clause

— It states conditions that the results must satisfy

— But refers to group values, not data-item values

• HAVING clause conditions may refer to:

— A column in the GROUP BY clause

— A built-in function for a selected column

Example:

SELECT DNO, JOB, COUNT(*) AS TOTAL

FROM EMP
GROUP BY DNO, JOB
HAVING COUNT(*) >= 2

-dno- - - -job total----

200 SALESREP 4
400 CLERK 2
100 ACCOUNTANT 2

1
(

:• •

7-12 SG 3605
 Grouping Functions

Full Query Specification

SELECT [ ALL I DISTINCT ] { * I select-list)

FROM table-specification-list
[ WHERE Boolean-expression ]
[ GROUP BY column-specification-list]
[ HAVING Boolean-expression ]
[ ORDER BY sort-specification-list

• SELECT statements may contain WHERE, GROUP BY, and HAVING clauses

— SELECT clause identifies columns

— FROM, WHERE, GROUP BY, and HAVING clauses identify the tables and rows

- WHERE clause can improve performance

• Order of execution

1. The FROM clause creates a conceptual table from the tables to be used.

2. If there is a WHERE clause, rows not meeting its requirements are discarded
from the conceptual table made in step 1.

3. If there is a GROUP BY clause, a conceptual grouped table is formed from the

previously derived rows.

4. If there is a HAVING clause, rows not meeting its requirements are discarded
from the grouped table made in step 3.

5. The SELECT clause is executed to determine which columns of the remaining

rows are to be used.

Example:

SELECT DNO
FROM EMP
WHERE JOB = 'CLERK'
GROUP BY DNO
HAVING COUNT(*) >= 2

-dno-
400

SG 3605 7-13
Grouping Functions

Exercise 7-1

Using the functionality you have learned so far, write and execute the SQL statements
that will retrieve the requested data from the EMP and DEPT tables.

1. What is the average annual salary per job in each department? Recall the SAL
column is monthly salary.
fl-c)
V C- -„3-0
0

2. Display the department number and number of employees, who receive salaries only
in the entire company.
L'”
-t' '

0 tj

7-14 SG 3605
 Grouping Functions

3. Compute the average, minimum, and maximum salaries of those employees who are
clerks or managers. Display the job titles and results.

4. Display the department number of departments that have more than one clerk.

JV i
L

SG 3605 7-15
8
Joining Tables
 Joining Tables

Module 8
Joining Tables

Module Objectives 8-3

Benefit Statement 8-3

Materials 8-3

Equijoins 8-4

Join Guidelines 8-6

Multi-Table Joins 8-7

Self-Joins 8-10

Self-Join Guidelines 8-11

How RDMS Joins Tables 8-13

Exercise 8-1 8-14

SG 3605 8-1
 Joining Tables

Module Objectives

Upon completion of this module, you should be able to

• Formulate queries to join data from one or more tables.

Benefit Statement

Thus far the queries you have created have dealt with one table. This module expands
your capabilities to involve more than one table. The data that you will need to access will
frequently be in more than one table.

Materials

• OS 1100 UDS Relational Data Management System (UDS RDMS 1100) and IPF SQL
Interface End Use Guide

• OS 1100 UDS Relational Data Management System (UDS RDMS 1100) SQL
Programming Reference Manual

SG 3605 8-3
Joining Tables

Joins

• A join combines some or all columns and rows from two or more tables

• Types of joins

- Equijoins

- Cartesian joins

Self joins

— Non-equijoins

Equijoins

SELECT columns-from-tables-in-from-clause
FROM table 1, table2,
WHERE table l.column = table2.column
AND other-conditions

• An equijoin links rows from tables based on the equality of a common attribute
(column)

- The columns on which the tables are joined must have the same data type

- If column names are the same, they must be unambiguously referenced as

table.column

• Figure 8-1 illustrates joining two tables based on the department number field (DNO)

EMP
DEPT

EMPNO ENAME DNO DNAME LOC

Figure 8-1. Joining EMP and DEPT Tables

8-4 SG 3605
 Joining Tables

Equijoin Examples

Example 1. For each employee display the employee name, department number,
name, and location.

SELECT ENAME, JOB, EMP.DNO, DNAME, LOC

FROM EMP, DEPT
WHERE EMP.DNO = DEPT.DNO

FOSTER SALESREP 200 SALES NEW YORK

BROWN CLERK 200 SALES NEW YORK
WOODWORTH MANAGER 200 SALES NEW YORK
ROCKWELL ACCOUNTANT 100 ACCOUNT ING ATLANTA
MART IN SALES REP 200 SALES NEW YORK
ADAMS SALESREP 200 SALES NEW YORK
GLASS ACCOUNTANT 100 ACCOUNT ING ATLANTA
TURNER CLERK 100 ACCOUNT ING ATLANTA
LAWSON MANAGER 100 ACCOUNTING ATLANTA
JOHNSON CLERK 400 DISTRIBUTION PRINCETON
WILLIAMS PRESIDENT 300 MARKETING DETROIT
SMITH CLERK 400 DISTRIBUTION PRINCETON
FORD SALES REP 200 SALES NEW YORK
HYDE MANAGER 400 DISTRIBUTION PRINCETON

Example 2.

SELECT ENAME, JOB, EMP.DNO, DNAME, LOC

FROM DEPT, EMP
WHERE EMP.DNO = DEPT.DNO

Note: The second example will produce the same results as the first.

SG 3605 8-5
Joining Tables

Join Guidelines

• The tables to be joined are specified in the FROM clause

• The WHERE clause specifies how to join the tables as well as any other selection
conditions as in single table queries.

• The "matching criteria" for the tables are called the join criteria

• More than one pair of columns may be used to specify the join condition between any
two tables

• Columns from either table may be named in the SELECT list

• Table.* is used to select all columns of the table specified

• If a column has the same name in both tables, it must be unambiguously referenced
as table-name.column-name in the SELECT or WHERE clause.

• The join is more efficient if the columns used to perform the join are indexed

• As many tables as desired may be joined together, although more than three may
cause performance slowdown

• If the WHERE clause does not use primary or secondary keys, the order of the tables
in the FROM list is 4xot important.
- Think of RDMS 1100/2200 as scanning the list from left to right

- Only the records "found" in the first table need to be joined with records in the
second and subsequent tables
Specify as the first table the one with the fewest expected "hits"

8-6 SG 3605
 Joining Tables

Multi-Table Joins 1 11

• If joining more than two tables, specify a join criterion for each pair of tables being
joined.

• For joining n tables, there should be n-1 join criteria in the WHERE clause.

Example Assume there is a third table named AUX_EMP that contains personnel
information on each employee, such as GRADE, and having a primary
key of EMPNO. If you want to display the employee's name, grade,
position, department number, and location, the information can be
collected from EMP, DEPT, and AUX EMP. The three tables could be
joined with a query such as the following:

SELECT ENAME, GRADE, JOB, EMP.DNO, DNAME, LOC

FROM EMP, AUX EMP, DEPT
WHERE EMP.DNO = DEPT.DNO
AND EMP.EMPNO = AUX EMP.EMPNO

ENAME GRADE JOB DNO DNAME LOC

HYDE 4 MANAGER 400 DISTRIBUTION PRINCETON

ADAMS 3 SALESREP 200 SALES NEW YORK
LAWSON 4 MANAGER 100 ACCOUNTING ATLANTA

Note: The WHERE clause specifies the join conditions. Only a part of the resulting
table is shown for the above SELECT.

SG 3605 8-7
Joining Tables

Cartesian Joins

If you do not specify a join criterion, then RDMS does not know how you want the rows of
one table matched with the rows of the second table. The table that results is a Cartesian
product; every row of the first table is matched with every row of the second. This
produces a large table and is usually not what was wanted.

Cartesian joins exist for operational completeness, but when you get one, it is usually an
accident. You probably forgot to specify the desired join condition. The resulting table
has as many rows as the product of the number of rows in each table.

Example: Modify the previous equijoin example, omitting the WHERE clause and
displaying both department number fields. The query would be:

SELECT ENAME, JOB, EMP.DNO AS E_DNO, DEPT.DNO AS D_DNO,

DNAME, LOC FROM EMP, DEPT

The table answering that query would have 14 x 4 or 56 rows. The top of the table would
appear as follows:

--ENAME--- ---JOB--- E_DN0 D_DNO ---DNAME--- ---LOC----

FOSTER SALESREP 200 100 ACCOUNTING ATLANTA
FOSTER SALESREP 200 200 SALES NEW YORK
FOSTER SALESREP 200 300 MARKETING DETROIT
FOSTER SALESREP 200 400 DISTRIBUTION PRINCETON
BROWN CLERK 200 100 ACCOUNTING ATLANTA
BROWN CLERK 200 200 SALES NEW YORK
BROWN CLERK 200 300 MARKETING DETROIT
BROWN CLERK 200 400 DISTRIBUTION PRINCETON
WOODWORTH MANAGER 200 100 ACCOUNTING ATLANTA
WOODWORTH MANAGER 200 200 SALES NEW YORK
WOODWORTH MANAGER 200 300 MARKETING DETROIT
WOODWORTH MANAGER 200 400 DISTRIBUTION PRINCETON
ROCKWELL ACCOUNTANT 100 100 ACCOUNTING ATLANTA

Note: There are four entries for each employee; one for each row of possible values
from the DEPT table.

8-8 SG 3605
 Joining Tables

Table Aliases

FROM table1 Wiest table2 alias2,

• Any table may be given another name or table alias

• Long, descriptive names can be shortened

• When joining a table with itself, it must be given aliases to clarify the column
references

Example 1

FROM customer_info ci, houses_available ha

WHERE ci.price > ha.price

Example 2

FROM EMP WORKER, EMP MANAGER

WHERE WORKER.MGR = MANAGER.EMPNO

EMP (WORKER) EMP (MANAGER)

EMPNO ENAME JOB MGR EMPNO ENAME JOB MGR

5010 FOSTER SALESREP 5234 5010 FOSTER SALESREP 5234
5146 BROWN CLERK 5234 5146 BROWN CLERK 5234
5234 WOODWORTH MANAGER 5784 5234 WOODWORTH MANAGER 5784
5237 ROCKWELL ACCOUNTANT 5743 5237 ROCKWELL ACCOUNTANT 5743
5437 MARTIN SALESREP 5234 5437 MARTIN SALESREP 5234
5469 ADAMS SALESREP 5234 5469 ADAMS SALESREP 5234
5630 GLASS ACCOUNTANT 5743 5630 GLASS ACCOUNTANT 5743
5702 TURNER CLERK 5743 5702 TURNER CLERK 5743
5743 LAWSON MANAGER 5784 5743 LAWSON MANAGER 5784
5765 JOHNSON CLERK 5984 5765 JOHNSON CLERK 5984
5784 WILLIAMS PRESIDENT 5784 WILLIAMS PRESIDENT
5896 SMITH CLERK 5984 5896 SMITH CLERK 5984
5942 FORD SALESREP 5234 5942 FORD SALESREP 5234
5984 HYDE MANAGER 5784 5984 HYDE MANAGER 5784

SG 3605 8-9
Joining Tables

Self-Joins

SELECT allasi.column, alias2.column,

FROM table aliasi, table alias2
WHERE aliasl.column = alias2.column
AND other-conditions

Joins a table to itself, comparing values in two different columns with the same data
type.

• The column names in the select-list must be qualified with the alias to distinguish
which to display since the column names in the tables are identical.

Example: Display the number, name, and job of each employee, and the number, name,
and job of the employee's manager.

SELECT WORKER.EMPNO, WORKER.ENAME, WORKER.JOB, WORKER.MGR,

MANAGER.EMPNO, MANAGER.ENAME, MANAGER.JOB
FROM EMP WORKER, EMP MANAGER
WHERE WORKER.MGR = MANAGER.EMPNO

MGR MGR MGR

EMPNO ENAME JOB MGR EMPNO ENAME JOB

5010 FOSTER SALESREP 5234 5234 WOODWORTH MANAGER

5146 BROWN CLERK 5234 5234 WOODWORTH MANAGER
5234 WOODWORTH MANAGER 5784 5784 WILLIAMS PRESIDENT
5237 ROCKWELL ACCOUNTANT 5743 5743 LAWSON MANAGER
5437 MARTIN SALESREP 5234 5234 WOODWORTH MANAGER
5469 ADAMS SALESREP 5234 5234 WOODWORTH MANAGER
5630 GLASS ACCOUNTANT 5743 5743 LAWSON MANAGER
5702 TURNER CLERK 5743 5743 LAWSON MANAGER
5743 LAWSON MANAGER 5784 5784 WILLIAMS PRESIDENT
5765 JOHNSON CLERK 5984 5984 HYDE MANAGER
5896 SMITH CLERK 5984 5984 HYDE MANAGER
5942 FORD SALESREP 5234 5234 WOODWORTH MANAGER
5984 HYDE MANAGER 5784 5784 WILLIAMS PRESIDENT

Note: The WORKER.MGR and MANAGER.EMPNO columns are displayed for

illustration of the matching column references.

8-10 SG 3605
 Joining Tables

Self-Join Guidelines

• A table can be joined with itself as if joining two separate tables

• Self-join is used to join one row in a table with another in the same table

• The columns specified in the join criterion must have the same data type

• The table must be given two aliases to identify which columns are coming from which
copy of the table

• In the previous example, note the following:

- Both the EMPNO and MGR columns contain the same type of information,
employee numbers as four numeric digits

- For each row in the EMP table, EMPNO is the employee number of the employee
named in the row

- For each row in the EMP table, MGR is the employee number of the employee's
manager

- To obtain the name of the employee's manager, the EMP table must be joined
with itself on the MGR column, comparing that to the EMPNO column, then
locating the ENAME associated with that employee number

SG 3605 8-11
Joining Tables

Non-Equijoins

• Most joins are based on the equality of values in the joined column; therefore they are
called equijoins.

• Non-equijoins are based on other comparison operators:

— <>

<=

- >=

- BETWEEN ... AND

Example: List the salary grade, name, salary, and job title of each employee. Order by
grade, and within grade by salary.

SELECT GRADE, ENAME, SAL, JOB

FROM EMP, SAL GRADE
WHERE SAL BETWEEN LOSAL AND HISAL
ORDER BY GRADE, SAL

GRADE ENAME SAL JOB

1 TURNER 900 CLERK

1 JOHNSON 975 CLERK
1 BROWN 1000 CLERK
1 SMITH 1050 CLERK
2 GLASS 1850 ACCOUNTANT
2 ROCKWELL 2175 ACCOUNTANT
3 ADAMS 2250 SALESREP
3 FORD 2375 SALESREP
3 MARTIN 2500 SALESREP
3 FOSTER 2700 SALESREP
4 HYDE 2900 MANAGER
4 WOODWORTH 3250 MANAGER
4 LAWSON 3400 MANAGER
5 WILLIAMS 5200 PRESIDENT

8-12 SG 3605
 Joining Tables

How RDMS Joins Tables

• Determine the cross product of the tables listed in the FROM clause

- Produces a conceptual table, the cross-product table

Contains all the columns from all the joined tables

Number of rows is the product of the number of rows in each table

- Three tables with 4 rows each means 4 X 4 X 4 = 64 rows

If one table is empty, resulting cross-product table is empty

• Apply the conditions of the WHERE clause to every row in the cross-product table

- Produces the reduced cross-product table

- Number of columns remains the same

- Number of rows is reduced

• Apply the query specification clause to the reduced cross-product table

— Use the conditions in the SELECT clause

— Report only columns requested

Note: If the query has multiple WHERE conditions, they are typically AND'ed
together. A common mistake is to OR the conditions, selecting far too many
rows.

SG 3605 8-13
Joining Tables

Exercise 8-1

Using the functionality you have learned so far, write and execute SQL statements that
will retrieve data from the EMP and DEPT tables to answer the following questions.

1. Which employees work in Atlanta?

Pirvl
rC) cr)
Dr.10
t;f• C.,
-
I, 0 c Pi;.4.77,,, A
i.4-• f \I 0

2. How many employees work in New York?

(4.' f4A 0

3. List the employees' names and the cities in which they work. Order the list by city.

4,•`'"

4. Retrieve the name and salary of employees who earn more than their managers.
47, 7- a

'tl*:D
, .

5. List the department number, location, and number of employees for every location.

(-1 .f2-4117.:
TA-;;cfr 1..jtV3
O
(--;,(2. • ) )

8-14 SG 3605
9
More Retrieval
Considerations
 More Retrieval Considerations

Module 9
More Retrieval Considerations

Module Objectives 9-3

Benefit Statement 9-3

Materials 9-3

Data Access Considerations 9-4

Indicator Variables 9-4
Indicator Variable Examples 9-5
Handling Tables with More Than 28 Columns (ACOB) 9-6

Reusing Cursors 9-7

Cursor Review 9-7
Describe Reusable 9-7
Host Program Variables in Cursors 9-7
Program Flow for Reusing Cursors 9-8
CLOSE Command 9-9
DROP CURSOR Command 9-10
OPEN ... USING Command 9-11
Example of Host Variables in Cursor 9-11

Lab 9-1 9-13

Sequential and Random Processing on a Cursor 9-14

Currency and Fetching 9-14
FETCH [ option ] Command 9-15
Introduction to Random Processing on a Cursor 9-16
DECLARE CURSOR Command 9-17
LOCATE Command 9-17
FETCH Command 9-18
Random Processing Example 9-19

Lab 9-2 9-20

Retrieval Efficiency Considerations 9-21

Singleton SELECT Command 9-21
Singleton SELECT Example 9-22
FETCH NEXT n Command 9-23
FETCH NEXT n Example 9-24
How RDMS Retrieves Data 9-25
Retrieval Using No Key 9-26
Retrieval Using A Secondary Key 9-27
Secondary Keys 9-28

SG 3605 9-1
More Retrieval Considerations

Accessing Data Using Keys 9-29

Exercise 9-1 9-30

9-2 SG 3605
 More Retrieval Considerations

Module Objectives

Upon completion of this module, you should be able to

• Describe and use any of the various programming or processing features that will
help solve retrieval problems, including:

Using indicator variables

Passing more than 28 values in an ACOB SQL command

- Processing forward or backward through a cursor

Random access of the database

- Reusable cursors

- Retrieving a single record

Retrieving multiple records from a single fetch

Efficiently accessing the database using primary or secondary keys

Benefit Statement

This module equips you with skills needed to solve retrieval problems. In addition, many
of the considerations covered will give you insights into how to make your retrieval
programs more efficient.

Note: This modules contains a large number of Unisys extensions to SQL '89. If your
code must be portable, you will avoid using the extensions.

Materials

• OS 1100 UDS Relational Data Management System (UDS RDMS 1100) SQL
Programming Reference Manual

• OS 1100 UDS Relational Data Management System (UDS RDMS 1100)

Administration Guide

SG 3605 9-3
More Retrieval Considerations

Data Access Considerations

Indicator Variables

host-varl [ INDICATOR ] ind-varl, host-var2 [ INDICATOR J ind-var2, . . .

• Use to determine two different conditions

— Tell whether or not the retrieved data value is NULL

— Tell if a retrieved CHARACTER data column has been truncated

Table 9-1. Indicator Variable Values

Value Meaning

Negative NULL data value

0 Data value is non-NULL and fits into receiving field
Positive True length of truncated data value

• Anywhere that a placeholder variable ($P) or embedded variable refers to a NULLS

ALLOWED column, an indicator variable must follow it.

— Indicator variables are required in a program accessing that column

— Indicator variables provide programmer control over handling NULL values.

• Declare the host program variable corresponding to an indicator placeholder variable

or the embedded variable that serves as the indicator variable as signed numeric

• COBOL picture clause

PIC S9(4) SIGN LEADING SEPARATE

9-4 SG 3605
 More Retrieval Considerations

Indicator Variable Examples

Requirement: To compute total compensation

ACOB ISQL Example

WORKING-STORAGE SECTION.
01 SAL-IND PIC S9(4) SIGN LEADING SEPARATE.
01 COMM-IND PIC S9(4) SIGN LEADING SEPARATE.

PROCEDURE DIVISION.

•
ENTER MASM 'ACOB$RDMR' USING
'FETCH all_comp INTO $P1, $P2 $P3, $P4 $P5
ERR-STAT, AUX-INFO,
EMPNO-OUT,
SAL-OUT,
SAL-IND, <
COMM-OUT,
COMM-IND. <
IF COMM-IND < 0 MOVE 0 TO COMM-OUT.
COMPUTE TOTAL-COMP = SAL-OUT + COMM-OUT.

UCOB ESQL Example

WORKING-STORAGE SECTION.
EXEC SQL BEGIN DECLARE SECTION END-EXEC.
01 SAL-IND PIC S9(4) SIGN LEADING SEPARATE.
01 COMM-IND PIC S9(4) SIGN LEADING SEPARATE.
EXEC SQL END DECLARE SECTION END-EXEC.
PROCEDURE DIVISION.

EXEC SQL
FETCH all comp INTO :EMPNO-OUT, :SAL-OUT :SAL-IND,
Z :COMM-IND
:COMM-UT
END-EXEC.
IF COMM-IND < 0 MOVE 0 TO COMM-OUT
COMPUTE TOTAL-COMP = SAL-OUT + COMM-OUT

SG 3605 9-5
More Retrieval Considerations

Handling Tables with More Than 28 Columns (ACOB)

ENTER MASM 'RSA$PARAM' USING list-of-program-variables

• ACOB limits the number of parameters on any ENTER MASM call to 31.

• RDMS uses the first three parameters for:

- The SQL command
Error status code

i1i ary information

• This leaves 28 data and indicator parameters on any call to ACOB$RDMR.

• If a table has more than 28 columns, use the subroutine RSA$PARAM

— Build a parameter list
— Then call ACOB$RDMR, to execute the command

• Pass the program variables in the same order as the corresponding host variables in
the SQL command.

Example:

ENTER MASM 'RSA$PARAM' USING PV1, PV2, PV31.

ENTER MASM 'RSA$PARAM' USING PV32, PV33,.. PV62.
STRING 'FETCH NEXT MY-CURSOR INTO ',
'($Pl, $P2,
' $P61, $P62);'
DELIMITED BY SIZE INTO RCOM.
ENTER MASM 'ACOB$RDMR' USING RCOM, ERR-STAT, AUX-INFO.

9-6 SG 3605
 More Retrieval Considerations

Reusing Cursors

Cursor Review
You may recall that cursors provide a way to deliver rows extracted from a table to your
program. The FETCH command already discussed retrieves one row of the cursor and
correlates the values with your program variables.

The cursor defined in the DECLARE CURSOR command remains in existence until the
end of the thread or until it is dropped. A cursor uses system resources - space in
RDMS's internal tables - so it helps to free this space as soon as possible.

Describe Reusable
Suppose you were writing a program that would first print all the employees in
Department 100, then all the employees in Department 200, and so on. There are two
ways you could do this:

- Declare a separate cursor for each department, and have multiple open cursors in
your program, or

— Declare a single cursor and reuse it, re-evaluating the query for each different
value of DNO

Declaring one cursor saves system resources and is more efficient. RDMS supports
multiple open cursors in a program, but you should save them for the occasions when they
are necessary.

Host Program Variables in Cursors

In the above example, it is possible to declare a cursor for the first department, print the
department information, then drop the cursor. You would have to re-declare the cursor
for each department. At least the departments to specify are already known.

But what if you needed to accept the desired department as input from a user? How is
that information passed to the cursor?

Host Program Variables!

Just as host program variables were used to pass retrieved information back to your
program in a FETCH, they can be used to pass information from your program to an SQL
command, such as OPEN CURSOR.

This capability is possible from ISQL, not ESQL. The ENTER MASM call makes the
correspondence between the placeholder variables and your host program variables.

SG 3605 9-7
More Retrieval Considerations

Program Flow for Reusing Cursors

Figure 9-1 illustrates the typical program flow of a program that reuses cursors.

DECLARE CURSOR

ACCEPT INPUT/
INITIALIZE PROG VAR.

OPEN CURSOR

FETCH

OTHER PROCESSING

CLOSE CURSOR (OPT)

END CONDITION?

DROP CURSOR (ISQL)

Figure 9-1. Program Flow for Reusing Cursors

9-8 SG 3605
 More Retrieval Considerations

CLOSE Command

CLOSE cursor-name

• Closes a previously opened cursor.

• Releases system space associated with the evaluation of the corresponding query with
current parameters

• Allows query to be re-evaluated with new parameters on next OPEN command

• Embedded (ESQL) - Interpreter (ISQL) interface differences

- CLOSE is optional for ISQL

- Consecutive OPENs invoke an implicit CLOSE for ISQL

In ESQL, an attempt to open an opened cursor results in an error

• RDMS automatically closes cursors if you execute a thread control command, END
THREAD for example, or if you encounter a rollback error

Example:

CLOSE HI_LO_SAL

SG 3605 9-9
More Retrieval Considerations

DROP CURSOR Command

DROP CURSOR cursor-name

• Unisys SQL extension for the interpreter interface (ISQL), not for ESQL

• Undoes the DECLARE CURSOR command

• Frees the system space used to define the cursor and its associated query

• References to a dropped cursor generate an error

Examples:

DROP CURSOR HI_LO_SAL

DROP CURSOR DEPT200

9-10 SG 3605
 More Retrieval Considerations

OPEN ... USING Command

OPEN cursor-name USING host-variable-list

• Unisys extension to for ISQL, the interpreter interface, not for ESQL

• Host variables used must correspond to those in DECLARE CURSOR

- $P1 in the DECLARE CURSOR must correspond to the $P1 in the OPEN

- DECLARE may use same variable more than once in WHERE clause

— OPEN references variable just once

— Order of host variables should be same as in DECLARE CURSOR for clarity

• Variables must be initialized (in working-storage or as input) before OPEN refers to

them

Query in DECLARE CURSOR is evaluated with values from OPEN call to determine
result

Open examples:

OPEN dept cursor USING $P1

OPEN hi_lo_sal USING $P1, $P2

Example of Host Variables in Cursor

Requirement: To read and display the name and employee number of employees who fall
in the salary range between hi-sal-in and lo-sal-in which are input by a user.

The following example illustrates the interpreter interface in a UCS COBOL program.
Recall that UCS COBOL supports both the interpreter interface (ISQL) as well as the
embedded interface (ESQL).

SG 3605 9-11
More Retrieval Considerations

WORKING-STORAGE SECTION.
01 HI-SAL-IN PIC.9(4) VALUE ZEROS.
01 LO-SAL-IN PIC 9(4) VALUE ZEROS.
01 RECORD-RETRIEVED
05 ENAME-OUT PIC X(10) VALUE SPACES.
05 EMPNO-OUT PIC 9(4) VALUE ZEROS.

PROCEDURE DIVISION.

MOVE 'DECLARE HI_LO_SAL CURSOR ' TO RCOM

MOVE ' SELECT ENAME, EMPNO FROM EMP TO RLIN(2)
MOVE ' WHERE SAL BETWEEN $P1 AND $P2 ;' TO RLIN(3)
ENTER MASM 'ACOB$RDMR' USING RCOM, ERR-STAT,
AUX-INFO, LO-SAL-IN, HI-SAL-IN

ACCEPT LO-SAL-IN
ACCEPT HI-SAL-IN
MOVE 'OPEN HI_L0 SAL USING $P1, $P2 ;' TO RCOM
ENTER MASM 'ACOB$RDMR' USING RCOM, ERR-STAT,
AUX-INFO, LO-SAL-IN, HI-SAL-IN

P▪ ERFORM UNTIL ERR-STAT = 6001

MOVE 'FETCH HI LO SAL INTO $P1, $P2 ;' TO RCOM
ENTER MASM 'ACOB$RDMR' USING RCOM, ERR-STAT,
'AUX-INFO, ENAME-OUT, EMPNO-OUT
DISPLAY RECORD-RETRIEVED
END-PERFORM

M▪ OVE 'CLOSE HI LO SAL ;' TO RCOM

ENTER MASM 'ACOB$RDMR' USING RCOM, ERR-STAT, AUX-INFO

M▪ OVE 'DROP CURSOR HI_LO SAL ;' TO RCOM

ENTER MASM 'ACOB$RDMR'GUIN RCOM, ERR-STAT, AUX-INFO

9-12 SG 3605
 More Retrieval Considerations

Lab 9-1

Reusing Cursors

Write and execute a COBOL program that uses only one cursor to do the following:

Retrieve and display the employee number," name, department number, and commission of
all employees except those in a specified department. If the commission is NULL, display
the word "NONE" in the commission column Accept the department as input and repeat
the display until a dummy department number of 999 is keyed in.

SG 3605 9-13
More Retrieval Considerations

Sequential and Random Processing on a Cursor

Currency and Fetching

• A cursor associates a named data area with a query

- Think of it as a "virtual table" that has currency

- Opening the cursor sets the pointer to just prior to the first row of the virtual
table.

• Currency allows processing through the cursor using different FETCH commands.

— FETCH FIRST

— FETCH [NEXT]

FETCH LAST

FETCH PRIOR

FETCH CURRENT

EMPNO ENAME

First >5984 HYDE

5469 ADAMS
5743 LAWSON
5765 JOHNSON
Prior >5630 GLASS Currency pointer
Next >5437 MARTIN
5784 WILLIAMS
5942 FORD
5702 TURNER
5234 WOODWORTH
5237 ROCKWELL
5010 FOSTER
5896 SMITH
Last >5146 BROWN

Figure 9-2. FETCH Options

9-14 SG 3605
 More Retrieval Considerations

FETCH [ option ] Command

FETCH [ option ] cursor-name INTO variable-specification-list

• Unisys SQL extension

• Where option may be FIRST, NEXT, LAST, PRIOR

- NEXT is default

- Cannot fetch LAST or PRIOR from a cursor declared with a GROUP BY,
HAVING, or ORDER BY clause

• The variable-specification-list is a list separated by commas consisting of:

— Placeholder variables for ISQL

— Embedded variables for ESQL

• Retrieves one row from the cursor relative to the current pointer position

• If the cursor is positioned after the last row and you execute a FETCH or FETCH
NEXT, or if you execute a FETCH PRIOR from before the first row of the cursor, an
end-of-cursor status is returned to you.

SG 3605 9-15
More Retrieval Considerations

Introduction to Random Processing on a Cursor

Unisys SQL extensions allow you to perform random processing on a cursor. The
commands that support this are not part of the ANSI standard SQL. Be aware of this if
you have concerns about the portability of your program.

Suppose you want to process different parts of a cursor separately. For example, from the
EMP table, display first the two rows starting with 5010, then the next two starting with
5784, then three more starting with 5469.

START 1
5010 FOSTER SALESREP
5146 BROWN CLERK
5234 WOODWORTH MANAGER
5237 ROCKWELL ACCOUNTANT
3 5437 MARTIN SALESREP
5469 ADAMS SALESREP
5630 GLASS ACCOUNTANT
5702 TURNER CLERK
5743 LAWSON MANAGER
2 5765 JOHNSON CLERK
5784 WILLIAMS PRESIDENT
5896 SMITH CLERK
5942 FORD SALESREP
5984 HYDE MANAGER

Figure 9-3. Random Processing

• To perform this type of random processing the steps are:

- DECLARE a cursor for random access

Set the currency pointer with a LOCATE command

FETCH the current row

- FETCH NEXT or PRIOR as desired

Loop back to the LOCATE command

9-16 SG 3605
 More Retrieval Considerations

DECLARE CURSOR Command

DECLARE cursor-name CURSOR

FOR RANDOM [ ACCESS ]
SELECT select-list
FROM table-name
[ WHERE condition(s)]

• Unisys extension to SQL

• SEQUENTIAL is the default

• A RANDOM ACCESS cursor can refer to only one table in the FROM clause

- Cannot contain FOR UPDATE, GROUP BY, or ORDER BY clauses

Cannot refer to a view or built-in function

- Cannot include the DISTINCT keyword

LOCATE Command

LOCATE cursor-name
ON table-name
USING VALUES ( primary-key-value )

• Unisys extension to SQL

• LOCATE uses the index of the primary key to position the currency pointer

- Is allowed on random access cursor only

Does not retrieve any data

SG 3605 9-17
More Retrieval Considerations

• The primary-key-value

- Is a list of values (separated by commas) of the column or cohimns of the table's

primary key

- Must be the same as when the table was created

If primary-key-value from LOCATE does not exist in table:

Auxiliary information variable is set to 1 (0 if it does exist)

Currency pointer is positioned to where key would be located if it existed, but

cursor does not have a current row

- FETCH CURRENT returns 6001 end-of-cursor status

FETCH NEXT returns next record after currency pointer

FETCH Command

FETCH CURRENT cursor-name

INTO host-variable-list

• Unisys extension to SQL

• FETCH CURRENT returns row to which currency pointer is set

9-18 SG 3605
 More Retrieval Considerations

Random Processing Example

Processing requirement:

From the EMP table, display first the two rows starting with 5010, then the next two
starting with 5784, then three more starting with 5469.

MOVE 'OPEN EMPS ; ' TO RCOM

ENTER MASM 'ACOB$RDMR' USING RCOM, ERR-STAT, AUX-INFO
MOVE 'LOCATE EMPS ON EMP ' TO RCOM
MOVE ' USING VALUES (5010) ;' TO RLIN(2)
ENTER MASM 'ACOB$RDMR' USING RCOM, ERR-STAT, AUX-INFO
MOVE 'FETCH CURRENT EMPS INTO $P1, $P2 ;' TO RCOM
ENTER MASM 'ACOB$RDMR' USING RCOM, ERR-STAT, AUX-INFO,
EMPNO-OUT, ENAME-OUT
DISPLAY RECORD-RETRIEVED
PERFORM FORWARD-READ

MOVE 'LOCATE EMPS ON EMP ' TO RCOM

MOVE ' USING VALUES (5784) ;' TO RLIN(2)
ENTER MASM 'ACOB$RDMR' USING RCOM, ERR-STAT, AUX-INFO
MOVE 'FETCH CURRENT EMPS INTO $P1, $P2 ;' TO RCOM
ENTER MASM 'ACOB$RDMR' USING RCOM, ERR-STAT, AUX-INFO,
EMPNO-OUT, ENAME-OUT
DISPLAY RECORD-RETRIEVED
PERFORM FORWARD-READ

MOVE 'LOCATE EMPS ON EMP ' TO RCOM

MOVE ' USING VALUES (5469) ;' TO RLIN(2)
ENTER MASM 'ACOB$RDMR' USING RCOM, ERR-STAT, AUX-INFO
MOVE 'FETCH CURRENT EMPS INTO $P1, $P2 ;' TO RCOM.
ENTER MASM 'ACOB$RDMR' USING RCOM, ERR-STAT, AUX-INFO,
EMPNO-OUT, ENAME-OUT
DISPLAY RECORD-RETRIEVED
PERFORM FORWARD-READ 2 TIMES
FORWARD-READ.
MOVE 'FETCH NEXT EMPS INTO $P1, $P2 ;' TO RCOM
ENTER MASM 'ACOB$RDMR' USING RCOM, ERR-STAT, AUX-INFO,
EMPNO-OUT, ENAME-OUT
DISPLAY RECORD-RETRIEVED.

SG 3605 9-19
More Retrieval Considerations

Lab 9-2

Write the PROCEDURE DIVISION that would read and display the rows of the DEPT
table in reverse DNO order.

9-20 SG 3605
 More Retrieval Considerations

Retrieval Efficiency Considerations

Singleton SELECT Command

SELECT [ ALL I DISTINCT ] { * I select-list }

INTO variable-specification-list --: -
FROM table-specification-list
[ WHERE Boolean-expression ]
[ GROUP BY column-specification-list ]
[ HAVING Boolean-expression ]

• Replaces DECLARE, OPEN, FETCH, and DROP with one command

- SELECT, FROM, WHERE, GROUP BY, and HAVING clauses are the same as for
the query specification

INTO clause is the same as for the FETCH command

• Returns one row from a table or tables to a program, typically using the primary key
value to guarantee only one row will be returned

• ISQL/ESQL differences

- For ISQL, if the WHERE clause does not identify a single row, the row retrieved
is the first one found

For ESQL, the singleton SELECT must result in 0 or 1 rows being selected, or
RDMS returns an error

SG 3605 9-21
More Retrieval Considerations

Singleton SELECT Example

Singleton SELECT:

ENTER MASM 'ACOB$RDMR' USING

'SELECT ename, sal
INTO $P1, $P2
FROM emp
WHERE empno = 5237 ,,
err-stat, aux-info, ename-out, sal-out

Replaces:

ENTER MASM 'ACOB$RDMR' USING

'DECLARE CURSOR my_cursor
SELECT ename, sal FROM emp
WHERE empno = 5237 ;',
err-stat, aux-info

ENTER MASM 'ACOB$RDMR' USING

'OPEN my_cursor
err-stat, aux-info

ENTER MASM 'ACOB$RDMR' USING

'FETCH my_cursor INTO $P1, $P2 ;',
err-stat, aux-info, ename-out, sal-out

9-22 SG 3605
 More Retrieval Considerations

FETCH NEXT n Command

FETCH NEXT number-of-rows [ FROM ] cursor-name

INTO variable-specification-list

• Use to fetch multiple rows from a cursor

• Available through the embedded interface (ESQL), not ISQL

• This command is not available in SQL 89

• The number-of-rows must be a positive integer constant

• The variable-specification-list is a list of embedded variables

- Each is a one-dimensional array, declared as an elementary item

Array size is determined by number-of-rows

• Status information returned by RDMS

— Auxiliary information is the number of rows returned

— Auxiliary information should always be checked after a FETCH NEXT n

— If fewer than n rows are fetched, error status = 6001 and SQLCODE = +100

SG 3605 9-23
More Retrieval Considerations

FETCH NEXT n Example

This example illustrates the PROCEDURE DIVISION and corresponding DATA

DIVISION items in a UCS COBOL program for fetching 5 rows of the employee table at a
time into an array. Only the column corresponding to RR-COMM allowed NULLs.

01 REC-RETRIEVED.
05 RR-EMPNO OCCURS 5 PIC 9(5).
05 RR-ENAME OCCURS 5 PIC X(10).
05 RR-JOB OCCURS 5 PIC X(10).
05 RR-MGR OCCURS 5 PIC 9(5).
05 RR-DNO OCCURS 5 PIC 9(4).
05 RR-HIREDATE OCCURS 5 PIC 9(7).
05 RR-SAL OCCURS 5 PIC 9(5).
05 RR-COMM OCCURS 5 PIC 9(5).
01 REC-COMM-IND.
05 RR-COMM-IND OCCURS 5 PIC S9(4)
SIGN LEADING SEPARATE.

EXEC SQL
DECLARE Cl CURSOR SELECT * FROM EMP
END-EXEC.

EXEC SQL
OPEN CURSOR Cl
END-EXEC.

EXEC SQL
FETCH NEXT 5 FROM Cl INTO :RR-EMPNO, :RR-ENAME,
:RR-JOB, :RR-MGR, :RR-DNO, :RR-HIREDATE,
:RR-SAL, :RR-COMM :RR-COMM-IND
END-EXEC.

9-24 SG 3605
 More Retrieval Considerations

How RDMS Retrieves Data

When programming for better performance, it helps to know how RDMS works and to
take advantage of its features. It is easy to write SQL queries and RDMS will eventually
return the rows requested no matter how inefficiently the query is written. But, there are
steps you can take to make it easier and more efficient to find the data. To be able to
write better queries, you need to know how RDMS locates the data you request.

Retrieval Using the Primary Key

This is the most efficient type of access. Tables are stored in primary key order and
RDMS can go directly to the page on which a row identified by primary key value is
located.

SELECT EMPNO, ENAME, DNO FROM EMP

WHERE EMPNO = 5743

Primary Index Data

Page 1 Page 2 P age 3
5437 1 5010 . . . 5469 ... 5784 ...
-> 5765 2 5146 . . . 5630 ... 5896 ...
5984 3 5234 5702 ... 5942 ...
5237 5743 ... 5984 ...
5437 ... 5765 ...

5010 FOS TER SALESREP 5234 200 • •

5146 BROWN CLERK 5234 200 ..

5234 WOODWORT H MANAGER 5784 200 • •

5237 ROCKWELL ACCOUNTANT 5743 100 ..

5437 MART IN SALESREP 5234 200 • •

5469 ADAMS SALE SREP 5234 200 ..

5630 GLASS ACCOUNTANT 5743 100 ..
5702 TURNER CLERK 5743 100 ..
> 5743 LAWSON MANAGER 5784 100 ..
5765 JOHNSON CLERK 5984 400 ..
5784 WILLIAMS PRES IDENT 300 • •

5896 SMITH CLERK 5984 400 ..

5942 FORD SALESREP 5234 200 • •

5984 HYDE MANAGER 5784 400 ..

Figure 9-4. Primary Key Usage

SG 3605 9-25
More Retrieval Considerations

Retrieval Using No Key

This is the slowest type of access. RDMS must walk the entire table, "looking at" each
row, and evaluating whether it satisfies the conditions of the query.

DNO has not been specified as a secondary key.

SELECT EMPNO, ENAME, DNO FROM EMP

WHERE DNO = 100

>5010 FOSTER SALESREP 5234 200 • •

>5146 BROWN CLERK 5234 200 • •

>5234 WOODWORTH MANAGER 5784 200 • •

>5237 ROCKWELL ACCOUNTANT 5743 100 • • • Satisfies

>5437 MARTIN SALE SREP 5234 200 • •

>5469 ADAMS SALE SREP 5234 200 •

>5630 GLASS ACCOUNTANT 5743 100 • • Satisfies

>5702 TURNER CLERK 5743 100 • • Satisfies
>5743 LAWSON MANAGER 5784 100 • • • Satisfies
>5765 JOHNSON CLERK 5984 400 • •

>2784 WILLIAMS PRESIDENT 300 •

>5896 SMITH CLERK 5984 400 • •

>5942 FORD SALESREP 5234 200

>5984 HYDE MANAGER 5784 400 • •

Figure 9-5. No Key Usage

9-26 SG 3605
 More Retrieval Considerations

Retrieval Using A Secondary Key

This is a fast type of access. RDMS has an index that lists the secondary key values and
the corresponding primary key. RDMS can then index into the data using the primary
key. If a column is used very frequently for access but is not the primary key, and the
processing requires fast response, the database designer may implement it as a secondary
key.

DNO has been specified as a secondary key.

SELECT EMPNO, ENAME, DNO FROM EMP

WHERE DNO = 100

Secondary Primary Data

Index Index Page 2 Page 3 Page 4

—> 100 5237 -> 5437 2 5010 ... 5469 ... 5784 ...
—> 100 5630 -> 5765 3 5146 ... 5630 ... 5896 ...
—> 100 5702 -> 5984 5234 ... 5702 ... 5942 ...
—> 100 5743 -> 5237 ... 5743... 5984 ...
200 5010 5437 5765 ...
200 5146
200 5234
200 5437
200 5469
200 5942
300 5784
400 5765
400 5896
400 5984

Figure 9-6. Secondary Key Usage

SG 3605 9-27
More Retrieval Considerations

Secondary Keys

• A secondary key is a set of one or more columns for which RDMS creates an index
that provides fast access to those columns

• A secondary index should be created when:

- Rows are frequently accessed using those columns
- The table has many rows

• If the EMP table were often queried using DNO and ENAME, an index for each
should be declared

DNO INDEX ENAME INDEX

DNO EMPNO ENAME EMPNO

100 5237 ADAMS 5469

100 5630 BROWN 5146
100 5702 FORD 5942
100 5743 FOSTER 5010
200 5010 GLASS 5630
200 5146 HYDE 5984
200 5234 JOHNSON 5765
200 5437 LAWSON 5743
200 5469 MART IN 5437
200 5942 ROCKWELL 5237
300 5784 SMITH 5896
400 5765 TURNER 5702
400 5896 WILLIAMS 5784
400 5984 WOODWORT H 5234

Figure 9-7. DNO and ENAME Indexes

9-28 SG 3605
 More Retrieval Considerations

Accessing Data Using Keys

• Performance is improved when the columns selected by your query are all part of the
primary or secondary key

Example:

SELECT ENAME, DNO

FROM EMP
WHERE DNO = 200

• When multiple secondary indices exist, RDMS uses the one declared first in the
CREATE TABLE command (other factors being equal)

Example:

SELECT ENAME, SAL

FROM EMP
WHERE DNO = 200
AND ENAME = 'FORD'

• If the key is multi-column, the index is used only if the WHERE clause uses the
leading column(s) of the index

SG 3605 9-29
More Retrieval Considerations

Exercise 9-1
Describe the programming or processing features that might help solve the following
retrieval problems.

1. Nothing returned when two fields were added together.

2. Trying to fetch 40 columns of information on a given fetch.

3. Problems encountered when multiple cursors were operative.

tj r
t.,C1r

4. Ability to easily retrieve employee records given their social security number.

9-30 SG 3605
 More Retrieval Considerations

5. An ability to retrieve a single employee record quickly.

6. Database programs that need to run much faster.

SG 3605 9-31
10
Data Manipulation
 Data Manipulation

Module 10
Data Manipulation

Module Objectives 10-3

Benefit Statement 10-3

Materials 10-3

Data Manipulation 10-4

BEGIN THREAD Command Revisited 10-5

ROLLBACK Command 10-6

Inserting Rows into Tables 10-7

Introduction to INSERT 10-7
INSERT Command 10-8

Changing Existing Data 10-9

Introduction to UPDATE 10-9
UPDATE Positioned Command 10-10
UPDATE Searched Command 10-11
Sample COBOL Update Program 10-12

Deleting Rows from Tables 10-16

DELETE Command 10-16
DELETE Examples 10-17
Sample COBOL Delete Program 10-18

Cursor Requirements for Commands 10-21

Exercise 10-1 10-22

Lab 10-1 10-23

SG 3605 10-1
 Data Manipulation

Module Objectives

• Change the data in your tables.

- Open a thread for update.

- Use the UPDATE, INSERT, and DELETE commands to modify data.

- Use the ROLLBACK command to reverse the effects of data changes.

Benefit Statement

You will see in this module how easy it is to modify your database with a few SQL
commands. You probably realize, having said that it is easy to change, that you will need
to deploy safeguards to guarantee the integrity of your database. A few of the integrity
capabilities are touched on here as well, but more on that later.

Note: There are several Unisys extensions for the UPDATE command and one for
the INSERT command. Those syntax extensions are not discussed in this
module, only the SQL 89 formats are presented. If you are maintaining or
migrating SQL code, consult the PRM listed below for the details on the
extensions.

Materials

• OS 1100 UDS Relational Data Management System (UDS RDMS 1100) and IPF SQL
Interface End Use Guide

• OS 1100 UDS Relational Data Management System (UDS RDMS 1100) SQL
Programming Reference Manual

SG 3605 10-3
Data Manipulation

Data Manipulation

• Data manipulation commands change the data residing in the data tables

- INSERT - Add new rows to a table

- UPDATE - Change values of columns of existing rows

DELETE - Remove rows from a table

• Query commands are often used to check the results of changes

Update command structure similar to query structure

If issued from an application program, DECLARE CURSOR is not always

required

Just MOVE update command to RCOM

Typically, you should use the BEGIN THREAD command

— Notify RDMS 1100 that you will update, not just read data

— Specify type of recovery desired

10-4 SG 3605
 Data Manipulation

BEGIN THREAD Command Revisited

BEGIN THREAD [ thread-name ]

FOR [ APPLICATION ] alias-name [ options]

• This is a Unisys extension

• For retrieval, the options discussed were:

[ READ I RETRIEVE ] [ UDSMSG UDSMESSAGE UDSMESSAGES

• For insertion, deletion, or update, the options are:

UPDATE( recovery-option) [ UDSMSG I UDSMESSAGE I UDSMESSAGES

• Choices for recovery-option are:

\ P
COMMANDLOOKS

- DEFERRED (the default)

- QUICKLOOKS

- NONE

Examples:

BEGIN THREAD FOR UDSSRC UPDATE(DEFERRED) UDSMSG

BEGIN THREAD T1 FOR APP1 UPDATE(QUICKLOOKS)

BEGIN THREAD FOR TEST1 UPDATE(DEFERRED) UDSMESSAGE

SG 3605 10-5
Data Manipulation

ROLLBACK Command

ROLLBACK WORK

• Before making any changes, it is important to know how to "undo" any modifications
in case of error

• ROLLBACK command discards any changes made to the database since the last
thread control command or rollback

• ROLLBACK will also

- Close all open cursors

Release all implicit locks on data or tables that were changed

• ROLLBACK is useful in application programs

When several updates are to be performed, but only if all of them can be
performed

- To do temporary updates to generate "what if' reports

10-6 SG 3605
 Data Manipulation

Inserting Rows into Tables

Introduction to INSERT

• Before inserting data into a table:

— The table must exist
— Data must satisfy any key and NULL value constraints

• Table need not initially contain any data

• INSERT command is used for interactive or small-scale row insertions

• Large-scale, initial load process is performed differently

Examples. Insert two new departments into the DEPT table as illustrated in Figure 10-1.

INSERT INTO DEPT

VALUES (500, 'RESEARCH', 'BOSTON')

INSERT INTO DEPT (DNAME, DNO, LOC)

VALUES ('OPERATIONS', 600, 'SALT LAKE')

DNO DNAME LOC

100 ACCOUNTING ATLANTA

200 SALES NEW YORK
300 MARKETING DETROIT
400 DISTRIBUTION PRINCETON

500 RESEARCH BOSTON

600 OPERATIONS SALT LAKE

Figure 10-1. Inserting New Departments

SG 3605 10-7
Data Manipulation

INSERT Command

INSERT INTO table-specification [ ( column-name-list)

{ VALUES ( value-list) query-specification

• Adds one or more rows to a new or existing table or view

- "Position" in table is determined by primary key value

- Duplicate primary key entry generates error message

• The column-name-list is used when:

- Inserting fewer columns than exist in the table definition

- Entering values in a sequence different than the table definition

If omitted, the value-list must correspond to the way the table was created, both
by order and number of columns

• The value-list provides values separated by commas supplied as

- Numeric constants or string literals

Host program variables ($Pn or embedded variables)

- Arithmetic expressions to be evaluated

- The keywords NULL or USER

• The query-specification inserts rows by selecting values from other tables

- Items in the SELECT list must correspond one to one, in order, with the
column-name-list

If the column-name-list is omitted, the items in the SELECT list must correspond
to the way the table was created, both by order and number of columns

INSERT INTO ROSVSCHEMM.DEPT

SELECT * FROM ROSVSCHEMA2.DEPT
WHERE DNO > 400

10-8 SG 3605
 Data Manipulation

Changing Existing Data

Introduction to UPDATE

• As long as primary key and foreign key constraints are obeyed, any data value in any
column or row can be changed.

- Change a primary key to another valid key value

- Change all the values in an entire column

Change a value in rows selected by a WHERE clause

• SQL 89 supports two forms of the UPDATE command

— Single row update

— Multiple row update

Note: Both forms of the command rely upon a WHERE clause to restrict the scope of
the update. If you omit the WHERE clause, all rows will be updated!

DNO DNAME LOC

100 ACCOUNTING ATLANTA

200 SALES NEW YORK
300 MARKETING DETROIT
400 DISTRIBUTION PRINCETON
Replace
500 RESEARCH BOSTON single
data value

600 OPERATIONS SALT LAKE

SEATTLE

> Replace entire row

700 PERSONNEL DENVER

Figure 10-2. Possible Updates

SG 3605 10-9
Data Manipulation

UPDATE Positioned Command

UPDATE table-specification
SET change-specification
WHERE CURRENT OF cursor-name

• Use to change the content of columns within the current row of the cursor

• The change-specification is a list of column names and values

- Form: column-name = value-specification [ ,column-name = value-specification ]

- Where value-specification may be a numeric constant, string literal, an arithmetic

expression, the keyword NULL, the keyword USER, or a variable specification

• Rules for referencing the cursor

- Cursor must be declared with the FOR UPDATE OF clause, a Unisys extension

- Only columns named in the FOR UPDATE OF clause can be updated

- Table names for DECLARE and UPDATE commands must match

- A successful FETCH must immediately precede the UPDATE

Example. Change Smith's job description from clerk to administrative assistant.

DECLARE UPDAT CURSOR

SELECT EMPNO, ENAME, JOB FROM EMP
WHERE EMPNO = 5896
FOR UPDATE OF JOB
FETCH UPDAT INTO :RR-EMPNO, :RR-ENAME, :RR-JOB
UPDATE EMP
SET JOB = 'ADMN ASST'
WHERE CURRENT OF UPDAT

SELECT * FROM EMP

WHERE EMPNO = 5896

EMPNO ENAME JOB MGR DNO HIREDATE SAL COMM

5896 SMITH ADMN ASST 5984 400 880913 1050

10-10 SG 3605
 Data Manipulation

UPDATE Searched Command

UPDATE table-specification
SET column-change-specification
[ WHERE Boolean-expression ]

• Change the content of columns within a subset of the existing rows of a table

• The change-specification is a list of column names and values

- Form: column-name = value-specification [ ,column-name = value-specification
Where value-specification may be a numeric constant, string literal, an arithmetic
expression, the keyword NULL, the keyword USER, or a variable specification

• The Boolean-expression specifies the rows to be updated

Example: Give all people in Department 200 who do not receive a commission a 7
percent increase in salary and a commission that is 10 percent of their
original salary.

UPDATE EMP
SET SAL = SAL * 1.07,
COMM = SAL * .10
WHERE DNO = 200
AND COMM IS NULL

SELECT * FROM EMP

WHERE DNO = 200

EMPNO ENAME JOB MGR DNO HIREDATE SAL COMM

5010 FOSTER SALESREP 5234 200 860714 2700 550

5146 BROWN CLERK 5234 200 871011 1070 100
5234 WOODWORTH MANAGER 5784 200 790219 3478 325
5437 MART IN SALESREP 5234 200 870326 2500 300
5469 ADAMS SALESREP 5234 200 811102 2250 500
5942 FORD SALESREP 5234 200 890316 2375 0

SG 3605 10-11
Data Manipulation

Sample COBOL Update Program

Requirement: Prompt and read data from a user to update an entire row in the EMP
table. Program is written in ISQL for UCS COBOL.
@UCOB'S DEV.MOD1O/UCOBUPDT,TPF$.UPDT,,,NO-OPTIONS,NARROW
UCOB- 5R2(910816) LSS- 6R1A(910814) 930419 13:37:59
1 IDENTIFICATION DIVISION.
2 PROGRAM-ID. update-emp.
3 ENVIRONMENT DIVISION.
4 CONFIGURATION SECTION.
5 SOURCE-COMPUTER. UNISYS-2200.
6 OBJECT-COMPUTER. UNISYS-2200.
7
8 DATA DIVISION.
9 WORKING-STORAGE SECTION.
10
11 01 thread-flag PIC 9 VALUE ZERO.
12 88 thread-has-begun VALUE 1.
13 01 err-end-flag PIC 9 VALUE ZERO.
14 88 no-more-msgs VALUE 1.
15
16 EXEC SQL BEGIN DECLARE SECTION END-EXEC.
17
18 01 update-record.
19 05 up-empno PIC 9(4).
20 05 up-ename PIC X(10).
21 05 up-job PIC X(10).
22 05 up-mgr PIC 9(4).
23 05 up-dno PIC 9(3).
24 05 up-hiredate PIC 9(6).
25 05 up-sal PIC 9(4).
26 05 up-comm PIC 9(4).
27 01 record-out.
28 05 empno-out PIC 9(4).
29 05 PIC X(2) VALUE SPACES.
30 05 ename-out PIC X(10).
31 05 PIC X(2) VALUE SPACES.
32 05 job-out PIC X(10).
33 05 PIC X(2) VALUE SPACES.
34 05 mgr-out PIC 9(4).
35 05 PIC X(2) VALUE SPACES.
36 05 dno-out PIC 9(3).
37 05 PIC X(2) VALUE SPACES.
38 05 hiredate-out PIC 9(6).
39 05 PIC X(2) VALUE SPACES.
40 05 sal-out PIC 9(4).
41 05 PIC X(2) VALUE SPACES.
42 05 comm-out PIC 9(4).
43

10-12 SG 3605
 Data Manipulation

44 01 SQLCODE PIC S9(9) USAGE BINARY.

45
46 01 RDMCA.
47 05 ERROR-STATUS PIC 9(4).
48 05 AUX-INFO PIC S9(9) USAGE BINARY.
49
50 01 err-text.
51 05 err-s PIC X(132) OCCURS 4 TIMES.
52
53 01 err-s-index PIC 9.
54 EXEC SQL END DECLARE SECTION END-EXEC.
55
56
57 PROCEDURE DIVISION.
58 000-main.
59
60 ***********************************************************
61 Begin thread
62 ***********************************************************
63
64 EXEC SQL
65 WHENEVER NOT FOUND GO TO 400-done
66 END-EXEC.
67
68 EXEC SQL
69 WHENEVER SQLERROR GO TO 900-rdms-error
70 END-EXEC.
71
72 BEGIN THREAD FOR UDSSRC UPDATE(DEFERRED).
73
74 ***********************************************************
75 Prompt and read input
76 ***********************************************************
77
78 DISPLAY 'Enter the current 4 digit employee number'
79 ACCEPT up-empno
80 DISPLAY 'Enter employee last name'
81 ACCEPT up-ename
82 DISPLAY 'Enter job title'
83 ACCEPT up-job
84 DISPLAY 'Enter employee number of manager'
85 ACCEPT up-mgr
86 DISPLAY 'Enter department number'
87 ACCEPT up-dno
88 DISPLAY 'Enter hiredate'
89 ACCEPT up-hiredate
90 DISPLAY 'Enter employee monthly salary'
91 ACCEPT up-sal
92 DISPLAY 'Enter commission, or NULL if not applicable'
93 ACCEPT up-comm
94

SG 3605 10-13
Data Manipulation

95 ***********************************************************
96 • Update row
97 ***********************************************************
98
99 EXEC SQL
100 UPDATE rosyschema2.emp
101 SET empno = :up-empno, ename = :up-ename, job = :up-job,
102 mgr = :up-mgr, dno = :up-dno, hiredate = :up-hiredate,
103 sal = :up-sal, comm = :up-comm
104 WHERE empno = :up-empno
105 END-EXEC.
106
107 DISPLAY 'Update Performed'.
108
109 ***********************************************************
110 • Display updated row
111 ***********************************************************
112
113 EXEC SQL
114 SELECT * INTO :empno-out, :ename-out, :job-out,
115 :mgr-out, :dno-out, :hiredate-out, :sal-out,
116 :comm-out FROM rosyschema2.emp
117 WHERE empno = :up-empno
118 END-EXEC.
119
120 DISPLAY 'The record was changed to:
121 DISPLAY record-out.
122
123 ***********************************************************
124 • End thread
125 ***********************************************************
126
127 400-done.
128
129 END THREAD.
130 MOVE 0 TO thread-flag.
131
132 DISPLAY 'Program has completed'.
133 STOP RUN.
134 **********************************************************
135 ** Do the RDMS error handling - display codes & messages
136 **********************************************************
137
138 *** Come here when RDMS errors are detected
139
140 900-rdms-error.
141
142 *** Turn off error handling, print error status variables
143
144 EXEC SQL WHENEVER SQLERROR CONTINUE END-EXEC.
145
146 DISPLAY '***SQL Code = ' SQLCODE.
147 DISPLAY '***Error Code = ' ERROR-STATUS.
148 DISPLAY '***Auxiliary Info = ' AUX-INFO.
149
150 *** Issue the error messages, end the thread, and exit
151
152 MOVE 0 TO err-end-flag.
153 PERFORM 940-geterror UNTIL no-more-msgs.
154 DISPLAY 'Error termination'.
155 END THREAD.
156 STOP RUN.

10-14 SG 3605
 Data Manipulation

157
158 *** Fill the error buffers using GETERROR
159
160 940-geterror.
161 GETERROR INTO :err-s(1), :err-s(2), :err-s(3), :err-s(4).
162 PERFORM 980-print-error
163 VARYING err-s-index FROM 1 BY 1 UNTIL err-s-index > 4.
164
165 *** Display the error buffers
166
167 980-print-error.
168 IF err-s(err-s-index) NOT = SPACE
169 DISPLAY err-s(err-s-index)
170 ELSE
171 MOVE 5 to err-s-index
172 MOVE 1 TO err-end-flag.
SIZES: CODE(EM4): 624 DATA: 1674
END UCOB- 0 ERRORS(MAJOR) 0 ERRORS(MINOR)
@XQT
Enter the current 4 digit employee number
Enter employee last name
Enter job title
Enter employee number of manager
Enter department number
Enter hiredate
Enter employee monthly salary
Enter commission, or NULL if not applicable
Update Performed
The record was changed to:
5469 ADAMS SALESREP 5234 200 870326 4000 1000

Program has completed

@BRKPT PRINTS

SG 3605 10-15
Data Manipulation

Deleting Rows from Tables

DELETE Command

DELETE FROM table-name

[ WHERE { Boolean-expression CURRENT OF cursor-name ]

• Deletes all rows satisfying the WHERE condition

- Boolean-expression specifies the row or rows to delete

CURRENT OF deletes the row where the cursor pointer is currently positioned

— All rows are deleted if there is no WHERE clause

• Rules for referencing the cursor

— Cursor cannot be read-only or for RANDOM ACCESS

— Table names must match for the DECLARE and DELETE commands

— A successful FETCH must immediately precede the DELETE command

— Cursor does not need to be declared with the FOR UPDATE OF clause

• RUMS returns the number of rows deleted through the auxiliary information
parameter

10-16 SG 3605
 Data Manipulation

DELETE Examples

Example 1. Remove the department where DNO = 200

DELETE FROM DEPT

WHERE DNO = 200

Example 2. Remove all employees in Department 200

DELETE FROM EMP

WHERE DNO = 200

Example 3. Empty the DEPT table

DELETE FROM DEPT

Example 4. Remove all employees named Smith

DELETE FROM EMP

WHERE ENAME = 'SMITH'

SG 3605 10-17
Data Manipulation

Sample COBOL Delete Program

Requirement: Prompt and read an employee number from a user and delete that row.
Continue prompting and deleting rows until a dummy value of 9999 is
read.
@ACOB,SE DEV.ACOBDELT,TPFS.REL
ACOB 7R1A R 75R3n 03/05/93 18:15:43 (0) 1100 ASCII COBOL 08/12/91 10:02
1100 ASCII COBOL SOURCE LISTING

1
2 IDENTIFICATION DIVISION.
3 PROGRAM-ID. TERMINATOR.
4 ENVIRONMENT DIVISION.
5 CONFIGURATION SECTION.
6 SOURCE-COMPUTER. UNISYS-2200.
7 OBJECT-COMPUTER. UNISYS-2200.
8
9 DATA DIVISION.
10 WORKING-STORAGE SECTION.
11 01 RCOM.
12 05 RLIN PIC X(40) OCCURS 12 TIMES.
13 01 ERR-STAT PIC 9(4).
14 01 AUX-INFO PIC S1(36).
15 01 ERROR-TEXT.
16 05 ERR-S PIC X(132) OCCURS 4 TIMES.
17 01 THREAD-FLAG PIC 9 VALUE ZERO.
18 88 THREAD-HAS-BEGUN VALUE 1.
19 01 D-EMPNO PIC 9(4).
20 *
21 PROCEDURE DIVISION.
22 ***************************************************************
23 Main Control Paragraph
24 ***************************************************************
25 000-MAIN.
26 PERFORM 200-INIT THRU 299-INIT-EX.
27 PERFORM 400-LOOP THRU 499-LOOP-EX
28 UNTIL D-EMPNO = 9999.
29 PERFORM 600 FIN THRU 699-FIN-EX.
30 STOP RUN.
31 ***************************************************************
32 Begin thread, read an employee number
33 ***************************************************************
34 200-INIT.
35 MOVE 'BEGIN THREAD FOR UDSSRC ' TO RCOM.
36 MOVE ' UPDATE (DEFERRED) • ' TO RLIN(2).
37 ENTER MASH 'acob$rdmr' USING RCOM, ERR-STAT, AUX-INFO.
38 IF ERR-STAT NOT = 0 GO TO 900-ERROR-PRINT.
39 MOVE SPACES TO RCOM.
40 MOVE 1 TO THREAD-FLAG.
41 DISPLAY 'Enter 4 digit employee number' UPON PRINTER.
42 ACCEPT D-EMPNO.
43 299-INIT-EX.
44 EXIT.

10-18 SG 3605
 Data Manipulation

45 **************************************************************
46 * Delete a row, solicit another
47 **************************************************************
48 400-LOOP.
49 MOVE 'DELETE FROM ROSVSCBEMA2.EMP ' TO RCOM.
50 MOVE 'WHERE EMPNO = $P1 ;' TO RLIN(2).
51 ENTER MASM 'ACOB$RDMR' USING RCOM, ERR-STAT,
52 AUX-INFO, D-EMPNO.
53 IF ERR-STAT NOT = 0 GO TO 900-ERROR-PRINT.
54 DISPLAY 'Record deleted' UPON PRINTER.
55 DISPLAY 'Enter 4 digit employee number' UPON PRINTER.
56 ACCEPT D-ENPNO.
57 499-LOOP-EX.
58 EXIT.
59 **************************************************************
60 • End the thread, issue all done
61 **************************************************************
62 600-FIN.
63 MOVE 'END THREAD;' TO RCOM.
64 ENTER MASK 'ACOB$RDMR' USING RCOM, ERR-STAT, AUX-INFO.
65 IF ERR-STAT NOT = 0000 PERFORM 900-ERROR-PRINT.
66 DISPLAY 'ALL DONE.' UPON PRINTER.
67 MOVE SPACES TO RCOM.
68 699-FIN-EX.
69 EXIT.
70 **************************************************************
71 • Print the error message, end the thread, shut it down *
72 **************************************************************
73 900-ERROR-PRINT.
74 DISPLAY 'RDML COMMAND = ' RCOM.
75 DISPLAY 'ERROR STATUS = ' ERR-STAT.
76 DISPLAY 'ERROR COLUMN = ' ADX-INFO.
77
78 MOVE 'GETERROR INTO $P1, $P2, $P3, $P4 ;' TO RCOM.
79 ENTER MASM 'ACOB$RDMR' USING RCOM, ERR-STAT, AUX-INFO,
80 ERR-S(1), ERR-S(2), ERR-S(3), ERR-S(4).
81 DISPLAY ERR-S(1) UPON PRINTER.
82 DISPLAY ERR-S(2) UPON PRINTER.
83 DISPLAY ERR-S(3) UPON PRINTER.
84 DISPLAY ERR-S(4) UPON PRINTER.
85
86 IF TBNEAD-BAS-BEGUN
87 MOVE 'END THREAD ;' TO RCOM
88 ENTER MASM 'ACOB$RDMR' USING RCOM,
89 ERR-STAT, ADX-INFO.
90
91 DISPLAY 'ERROR END' UPON PRINTER.
92 STOP RUN.
END ACOB DIAGNOSTIC TOTALS 0 WARNING 0 MINOR 0 SERIOUS 0
FATAL 0 LEVELING
COMPILE TIME IS 4 33 SECONDS CORE: MIN= 63601 / 64000 MAX* 66560 WORDS

SG 3605 10-19
Data Manipulation

@MAP,S ,TPF$.ABS
Collector 33R1A (910808 1521:53) 1993 Mar 05 Fri 1815:44
1. NOT TPF$.
2. IN TPFS.REL
3. IN SYS$LIBS*FtSA.CBEP$$RSA
4. IN SYS$LIB$*RSA.RDI/R-ACOBDAT
5. END
AFCM status of output element is SETAFCM
Quarter-word sensitive
ADDRESS LIMITS 001000 001726 471 IBANK WORDS DECIMAL
040000 061623 9108 DBANK WORDS DECIMAL
STARTING ADDRESS 001336

SEGMENT $MAIN$ 001000 001726 040000 061623

ERWRLIBT12 24
MAY 90 10:37:08
M$PKT$ 040000 040005 03
AUG 90 16:10:55
MEM$ERR
C$NPADS 040006 040014 26
AUG 90 19:05:18
C$S400 040015 040027 21
DEC 90 13:37:46
C$RETO 001000 001005 26
AUG 90 18:58:21
MEWERR(COMMONBLOCK) 040030 040032
C$5300 $(0) 040033 040730 25
JAN 91 11:32:50
$(012) MEM$ERR
C$PCRLOG $(0) 040731 041030 12
AUG 91 09:47:08
CEEP$$ACOB/PART-LIB-CB 12
AUG 91 10:13:59
REL $(1) 001006 001350 041031 041431 05
MAR 93 18:15:44
$ (3) 001351 001355 041432 041567
041570 041606
041607 041746
CBEP$$RSA 24
AUG 91 01:01:32
RDMR-ACOBDAT 001356 001726 $(0) 041747 061622 10
JAN 92 13:23:24
$(2) 061623 061623
Common Banks referenced
----------
C$S30S 0400103 C$S115 0400105 RDMR$FTNCONT 0401074
END MAP. ERRORS: 0 TIME: 23.502 STORAGE: 016412/015204/016170/0/0130076
@XQT
Enter 4 digit employee number
Record deleted
Enter 4 digit employee number
ALL DONE.

10-20 SG 3605
 Data Manipulation

Cursor Requirements for Commands

Table 10-1. Command Summary

Command Cursor Requirements

INSERT None required
UPDATE DECLARE of cursor must be with the
Positioned FOR UPDATE OF clause

Only columns named in the FOR

UPDATE OF clause can be updated

Cursor cannot be declared read-only or

for RANDOM ACCESS

Successful FETCH must immediately

precede the UPDATE
UPDATE None required
Searched
DELETE Cursor cannot be declared read-only or
for RANDOM ACCESS

Successful FETCH must immediately

precede the DELETE

Cursor does not need to be declared with

the FOR UPDATE clause

SG 3605 10-21
Data Manipulation

Exercise 10-1

Write the SQL code to perform the following data manipulation.

1. Change the salary for the employee with EMPNO = 5146 to 1250.

2. Change the commission for each salesman to 500.

Vc•-..1.
Q

3. Change EMPNO 5437 to 5620.

flo

4. Give all managers a 12 percent raise.

10-22 SG 3605
 Data Manipulation

Lab 10-1

Write and execute a COBOL program that reads values for a row of the DEPT table and
insert the row into the table. Have the program display the inserted row.

Extension: Continue to read and insert rows until a dummy value of 999 is entered for
DNO.

SG 3605 10-23
11
Keys and Updating
 Keys and Updating

Module 11
Keys and Updating

Module Objectives 11-3

Benefit Statement 11-3

Materials 11-3

Primary Keys 11-4

Update Anomalies 11-5

Foreign Keys 11-6

Introduction 11-6
Parent Table / Child Table Relationships 11-6
Foreign Key Considerations 11-7
Foreign Keys and INSERT 11-8
Foreign Keys and DELETE 11-9
Foreign Keys and UPDATE 11-10
Rules for Parent - Child Relationships 11-11

Lab 11-1 11-12

SG 3605 11-1
 Keys and Updating

Module Objectives

Upon completion of this module, you should be able to

• Update tables containing foreign keys.

— Define primary keys and foreign keys.

— Describe possible update anomalies and a method of maintaining referential

integrity.

— List the effects of foreign keys on deletions and updates.

Benefit Statement

How can you maintain consistency between tables? This is the main question answered in
this module. And once you establish the mechanisms to maintain consistency, how do you
update your database?

Materials

• OS 1100 UDS Relational Data Management System (UDS RDMS 1100)

Administration Guide

SG 3605 11-3
Keys and Updating

Primary Keys

• Unique identifier of each row

— One column or combination of columns
— Multi-column key called "composite key"

• Primary keys are specified when table is created

- Table is indexed on primary key
- Columns must be defined as NOT NULL

• Other columns may be specified as secondary indexes

- Advantage = Speeds up retrieval
Disadvantage = Adds overhead on updates

INDEX DNO DNAME LOC

100 100 ACCOUNT ING ATLANTA

200 200 SALES NEW YORK
300 300 MARKETING DETROIT
400 400 DISTRIBUTION PRINCETON

Figure 11-1. Primary Keys Illustrated

11-4 SG 3605
 Keys and Updating

Update Anomalies

• Anomaly - Deviation from the expected

• What would happen to the EMP table below if you executed the following updates?

UPDATE DEPT
SET DNO = 500
WHERE DNO = 200

INSERT INTO EMP

VALUES (5999,'GOOVSALESREP',523000,900701,2300,NULL)

• DNO is primary key of DEPT, but is also a column of EMP

• Possibility of update anomalies can be reduced by good database design

• RDMS can perform some integrity checking for you ES

Table 11-1. EMP

EMP NO ENAME JOB MGR I DNO I HIREDATE I SAL COMM

5010 FOSTER SALESREP 5234 200 860714 2700 400
5146 BROWN CLERK 5234 200 871011 1000
5234 WOODWORTH MANAGER 5784 200 790219 3250
5237 ROCKWELL ACCOUNTANT 5743 100 840618 2175
5437 MARTIN SALESREP 5234 200 870326 2500 300
5469 ADAMS SALESREP 5234 200 811102 2250 500
5630 GLASS ACCOUNTANT 5743 100 871220 1850
5702 TURNER CLERK 5743 100 890515 900
5743 LAWSON MANAGER 5784 100 830510 3400
5765 JOHNSON CLERK 5984 400 891221 975
5784 WILLIAMS PRESIDENT 300 781015 5200
5896 SMITH CLERK 5984 400 880913 1050
5942 FORD SALESREP 5234 200 890316 2375 0
5984 HYDE MANAGER 5784 400 850901 2900

SG 3605 11-5
Keys and Updating

Foreign Keys

Introduction

• A foreign key is a column in a table that is a primary key in another table

• Used to maintain data consistency between tables

• Defines a relationship between tables

• Causes data checking when tables are updated

• May be declared in CREATE TABLE command by DBA or added later

CREATE TABLE EMP

IN MY AREA
COLUMNS ARE EMPNO : DECIMAL(4),
ENAME : CHARACTER(1 0), ...
PRIMARY KEY KEY1 IS EMPNO ASCENDING
FOREIGN KEY KEY2 IS DNO
REFERS TO DNO OF DEPT

Parent Table I Child Table Relationships

• Table in which column is primary key is parent table

• Table in which column is foreign key is child table

• Figure 11-2 illustrates the parent - child relationship.

11-6 SG 3605
 Keys and Updating

PARENT
DEPT DNO DNAME LOC

100 ACCOUNTING ATLANTA

200 SALES NEW YORK
300 MARKETING DETROIT
400 DISTRIBUTION PRINCETON

CHILD
EMP

EMPNO ENAME JOB MGR DNO HIREDATE SAL COMM

5437 MART IN SALESREP 5234 200 870326 2500 300

5469 ADAMS SALESREP 5234 200 811102 2250 500
5630 GLASS ACCOUNTANT 5743 100 871220 1850
5743 LAWSON MANAGER 5784 100 830510 3400
5765 JOHNSON CLERK 5984 400 891221 975
5784 WILLIAMS PRES IDENT 300 781015 5200
5942 FORD SALESREP 5234 200 890316 2375 0
5984 HYDE MANAGER 5784 400 850901 2900

Figure 11-2. DNO in the Parent and Child Tables

Foreign Key Considerations

• One table can be parent of up to 50 children

• One table can be child of up to 25 parents

• A table may be parent of one foreign key and child of another

• Column-names of foreign key-primary key may be different, but datatypes of the

columns must be the same

• If foreign key relationship is added after tables have data, RDMS does not check
retroactively if data satisfy integrity constraints

SG 3605 11-7
Keys and Updating

Foreign Keys and INSERT

• Insert a row into child table

- Parent table must have a row with the primary key value corresponding to the
foreign key

- Then insert row in child table

• Sample problem

— If foreign key DNO is defined as a foreign key, as shown in Figure 11-2, and the
following INSERT is executed:

INSERT INTO EMP

VALUES (5999:GOOD',ZALESREP',5234,600,900701,2300)

— An error message is generated, since there is no DNO = 600

• Problem solution: Add the corresponding department to the DEPT table, then insert
the row in EMP

11-8 SG 3605
 Keys and Updating

Foreign Keys and DELETE

• Delete row in parent table

- There must be no rows in child table having primary key value from parent as
the foreign key

- Change rows in child table so foreign key is a valid primary key value

- Or, delete rows in child table

• Sample problem

— Attempt to delete a row in the DEPT table

DELETE FROM DEPT

WHERE DNO = 200

— Generates an error message since all the rows in EMP having DNO = 200 will be
stranded

• Solution to problem

— Delete all rows in EMP with DNO = 200

— Then delete row from DEPT where DNO = 200

SG 3605 11-9
Keys and Updating

Foreign Keys and UPDATE

• Update primary key value in parent table

There must be no rows in child table having old primary key value from parent
as the foreign key

- Change rows in hild table so foreign key is a valid primary key value

- Or, delete rows in child table

• Sample Problem: The following UPDATE generates an error message since all the
rows in EMP having DNO = 200 will be stranded

UPDATE DEPT
SET DNO = 500
WHERE DNO = 200

• Solution to above problem

- Insert a row in DEPT with DNO = 500

- Update the DNO of employees to 500 where DNO = 200

- Delete row from DEPT where DNO = 200

• Update foreign key value in child table

— May be changed to any other primary key value from parent table

Example: Transfers Brown from Department 200, Sales, to 400, Distribution

UPDATE EMP
SET DNO = 400
WHERE EMPNO = 5146

11-10 SG 3605
 Keys and Updating

Rules for Parent - Child Relationships

Cannot create children without parents

Cannot leave an orphan

Figure 11-3. Summary of Rules

SG 3605 11-11
Keys and Updating

Lab 11-1

Use IPF SQL or MAPPER to perform the following updates.

The tables you have been working with have a foreign key. The column DNO in EMP is
declared as a foreign key referring to DNO in DEPT. You will not use these tables again
after this exercise. You may commit or rollback the changes you make.

1. Try to insert a row with a non-existent DNO into EMP_

2. Perform the correct procedure for inserting the same row as above into EMP.

3. Try to delete a row directly from DEPT whose primary key is associated with one
or more rows in EMP,' '

4. Perform the procedure described in number 3 correctly.

5. Try to update a row in EMP, changing the DNO to a currently non-existent value.

6. Update a row as above, changing DNO from one its current value to another valid
value.

7. Try to change a DNO in DEPT which has a corresponding foreign key value in
EMP.

8. Perform the steps necessary to accomplish number 7 correctly.

11-12 SG 3605
12
Obtaining Table
Definition Information
 Obtaining Table Definition Information

Module 12
Obtaining Table Definition Information

Module Objectives 12-3

Benefit Statement 12-3

Materials 12-3

MAPPER RDI Access 12-4

Unisys Repository Manager (UREP 1100) 12-6

Introduction 12-6
Calling the Repository Manager 12-7
UREP Example 12-8

Lab 12-1 12-10

SG 3605 12-1
 Obtaining Table Definition Information

Module Objectives

Upon completion of this module, you should be able to

• Obtain table definition information in one of two ways:

- Using the MAPPER MRI interface

Using UREP

Benefit Statement

You will need to obtain table information to discover the data types and sizes for the
columns in tables. This information is needed for writing host program variables. The
table information provided by UREP will also tell you about primary, secondary, and
foreign keys that exist for your tables.

Materials

• MAPPER Relational Interface (MRI) Relational Database Interface (RDI) Operations

Guide

• OS 1100 UDS Unisys Repository Manager (UDS UREP) Programming Reference

Manual

SG 3605 12-3
Obtaining Table Definition Information

MAPPER RDI Access

How to obtain table column information using the MAPPER RDI interface.

Step 1. On the RDI Menu, transmit from the Select entry as shown in Figure 12-1.

RDI
RDI menu
Tab to your choice and press XMIT
Select
Join
Insert ►
Delete ►
Update ►
Create ►
SQL ►
Utilities menu
Detailed help ►
Rev. Level 3R3

3 4 5 6DBMS 7Envmnt 8Help 9 18Quit

Figure 12-1. RDI Menu

Step 2. On the Select menu, enter the table name and a question mark in the Column
names field. Notice that the table name is qualified by the name ROSVSCREMAL

RN
Select menu
Retrieve rows from a single relational table.
Table name rosyschemai.emp
Column names -> ?
Where clause ->
Retrieve all rows? -> y If no, how many? -> 1
Number of rows to skip -> Null Symbol -> *
Extra options? -> n
Place cursor here and press XMIT ->
Press XMIT from here for detailed help ->

1 2 3 4RDIMnu 5 6Specl 7 8Help 9 18Quit

Figure 12-2. Select Menu

12-4 SG 3605
 Obtaining Table Definition Information

Step 3. The RESULT displayed contains the column information for the EMP table, as
shown in Figure 12-3.

Line, 8 Roll, - RESULT

. DATE 83 MAR 93 09:38:39 REPORT GENERATION DAA
. Table 1 - rosyschemal.emp
. To select column names, enter the numbers 1-126 in the first field tit)
. and press XMIT. The columns will be displayed from left to right
. In the order you specify. Press the Resume key to continue. ,
* it .TAB. COLUMN NAME . TYPE .SIZE .DECPT.NF.PH
to=
1 EMPNO DECIMAL 5. ONY
1 ENAME CHARACTER 18 N N
1 JOB CHARACTER 10 V N
1 MGR DECIMAL 5 OYN
1 DNO DECIMAL 4 0 Y N
1 HIREDATE DECIMAL 7 0 Y N
1 SAL DECIMAL 5 8 V Pi
1 COMM DECIMAL 5 8 V N

.SIZE = size of column indicated by COLUMN NAME

.DECPT = number of decimal places if TYPE = decimal
.NF = null flag. N = nulls not allowed, Y = nulls allowed
.PH mary key,
= primary key. Y =_EPriREPORT , N = not primary key
D
1Resume Mint 3 4Return 5 6LineCh 7 8 10Quit

Figure 12-3. Result with Table Column Information

• Column information includes:

Column number

Number of the table being displayed

Column name

- Data type of the column

- Number of characters or size

Number of decimal places, if allowed

- If nulls area allowed

If the column is the primary key or a part of the primary key

SG 3605 12-5
Obtaining Table Definition Information

Unisys Repository Manager (UREP 1100)

Introduction

• Provides a focal point for management of corporate information resources

• UREP manages the repository, a meta-database, a relational database

• Benefits

— Locate information
- Identify how information is used

— Identify who uses information

- See how changes affect other related information

• When tables or views are created, information is written to the database

- Column names, data types, and sizes

- Keynames and composition

- Foreign key relationships

• Reports are available to provide information on how the tables and views are defined

12-6 SG 3605
 Obtaining Table Definition Information

Calling the Repository Manager

@DD,option source-input-element,source-output-element,application-group

• Some of the processor options include:

E echoing the incoming processor commands; they are not printed otherwise

D demand execution

B batch execution

X terminating UREP in error mode if executing in batch and an error is

encountered

• The source-input-element optionally specifies UREP commands

• The source-output-element optionally contains the output

• The application-group specifies the application group containing your tables

— If omitted, the default UDS Control application is used

— The UREP standard release default is UDSSRC

• Commands are written in the Data Dictionary System (DDS) command language

Note: Consult OS 1100 UDS Unisys Repository Manager (UDS UREP)

Programming Reference Manual for details on the commands.

SG 3605 12-7
Obtaining Table Definition Information

UREP Example
The following example illustrates the commands to obtain table information for the EMP
and DEPT tables.

@DD,E
UREP 1R2 (08/16/91 11:09:26) 03/26/93 15:48:14
*REMARK* UREP7403 The default UDS application ( UDSSRC ) will be used.
1 process table emp for schema rosyschemal report
** Report for Table EMP for Schema ROSVSCHEMA1 ***
Logical Filename: EMPLOYEE for Schema ROSVSCHEMA1
Preamble Length : 14
Rdt Length : 52
Total Length : 66
Owner Userid :
Data Access Control : ACTIVE
Base Relation Code : 1
Record Size(words) : 13
Number of Non-deleted Items : 8
Last Non-deleted Item : 8
Number of Foreign Keys : 0
Number of Member Relations : 0
Loading Order : RANDOM
Item Mod Flag : FALSE
Current Number Indexes : 0
Matching Order Flag : TRUE
Screen Type : COLUMNAR
Columns : Scale /
Name Code Type Size Precision
Nulls
EMPNO 1 DECIMAL 5 0
NOT-ALLOWED
ENAME 2 CHARACTER 10 0
NOT-ALLOWED
JOB 3 CHARACTER 10 0
ALLOWED
MGR 4 DECIMAL 0
ALLOWED
DNO 5 DECIMAL 4 0
ALLOWED
HIREDATE 6 DECIMAL 7 0
ALLOWED
SAL 7 DECIMAL 5 0
ALLOWED
COMM 8 DECIMAL 5 0
ALLOWED
Primary Key : ENO
Column, Direction
EMPNO, ASCENDING

12-8 SG 3605
 Obtaining Table Definition Information

2 process table dept for schema rosyschemal report

*** Report for Table DEPT for Schema ROSVSCHEMA1 ***
Logical Filename: DEPARTMENT for Schema ROSVSCHEMA1
Preamble Length : 14
Rdt Length : 41
Total Length : 55
Owner Userid :
Data Access Control : ACTIVE
Base Relation Code : 1
Record Size (words) : 7
Number of Non-deleted Items : 3
Last Non-deleted Item : 3
Number of Foreign Keys : 0
Number of Member Relations : 1
Loading Order : RANDOM
Item Mod Flag : FALSE
Current Number Indexes : 0
Matching Order Flag : TRUE
Screen Type : COLUMNAR
Columns : Scale /
Name Code Type Size Precision
Nulls
DNO 1 DECIMAL 4 0
NOT-ALLOWED
DNAME 2 CHARACTER 12 0
NOT-ALLOWED
LOC 3 CHARACTER 10 0
ALLOWED
Primary Key : K1
Column, Direction
DNO, ASCENDING
Member Relations (foreign key #)
ROSVSCHEMAl. EMP FOR KEY( 1)
END UREP ( 0 )FATALS ( 0 }ERRORS ( 0 }WARNINGS ( 1 )REMARKS .

SG 3605 12-9
Obtaining Table Definition Information

Lab 12-1

Use the MAPPER MRI interface or UREP directly to obtain table information for the
EMP table you are working with. How does the information received compare to the
information contained in this module?

12-10 SG 3605
13
Views
 Views

Module 13
Views

Module Objectives 13-3

Benefit Statement 13-3

Materials 13-3

Views 13-4
Basic Definition 13-4
What is a View? 13-5
Purpose of Views 13-6

CREATE VIEW Command 13-7

View Example 13-8
Columns in Views 13-9
Data Checking 13-10
Security 13-10

Updating Views 13-11

DROP VIEW Command 13-12

Exercise 13-1 13-13

Lab 13-1 13-14

SG 3605 13-1
 Views

Module Objectives

Upon completion of this module, you should be able to

• Create and use views

- Define a view and identify the advantages of views

- Identify the limitations of updates allowed on views

- Drop views

Benefit Statement

By working with views you will be able to solve several database problems. You will see
how to limit the information that is viewed or that can be updated by users. You also will
be able to safeguard your table from updates with "unrealistic" values.

Materials

• OS 1100 UDS Relational Data Management System (UDS RDMS 1100) and IPF SQL
Interface End Use Guide

• OS 1100 UDS Relational Data Management System (UDS RDMS 1100) SQL
Programming Reference Manual

SG 3605 13-3
Views

Views

Basic Definition
A view is a window into the data. Views provide multiple perspectives of the same
underlying tables.

Figure 13-1 shows three columns and a few rows that comprise the EMPMGR view.

EMPMGR View
EMPNO ENAME JOB
EMP Table
5010 FOSTER SALESREPT
5146 BROWN CLERK MGR DNO HIREDATE SAL COMM
5234 WOODWORTH MANAGER
5237 ROCKWELL ACCOUNTANT 5234 200 860714 2700 550
5234 200 871011 1000
5784 200 790219 3250
5743 100 840618 2175
5437 MARTIN SALESREP 5234 200 870326 2500 300
5469 ADAMS SALESREP 523 4 200 811102 2250 500
5630 GLASS ACCOUNTANT 5743 100 871220 1850
5702 TURNER CLERK 5743 100 890515 900
5743 LAWSON MANAGER 5784 100 830510 3400
5765 JOHNSON CLERK 5984 400 891221 975
5784 WILLIAMS PRESIDENT 300 781015 5200
5896 SMITH CLERK 5984 400 880913 1050
5942 FORD SALESREP 5234 200 890316 2375 0
5986 HYDE MANAGER 5784 400 850901 2900

Figure 13-1. Picturing a View

13-4 SG 3605
 Views

What is a View?

• A view is:

- A virtual table derived from underlying base tables

- Defined with a query specification (SELECT statement)

- Stored in the repository

- A direct window to the data in the tables

Referenced in the same way as a table

Qualifier. View-name. Column-name

• A view can present:

- A subset of the data in a table

- Some of the columns

Some of the rows

- A join of several tables

• A view is based on logical, not physical tables

• A view does not have its own data storage

SG 3605 13-5
Views

Purpose of Views

• Hides underlying database structure

• Simplifies data access (SELECT clause)

— Pre-selects columns

— Joins tables

• Masks unnecessary data

• Restricts access to confidential information

• Prevents users from inserting bad data

13-6 SG 3605
 Views

CREATE VIEW Command

CREATE VIEW [ Qualifier. ]view-name

[ ( column1, column2, column3... )]
AS query-specification - -
[WITH CHECK OPTION ]
[ WITH ACCESS CONTROL ]

• Unisys supports the SQL 89 form of this commmand; the WITH ACCESS CONTROL
clause is a Unisys extension to the standard command

• Qualifier defaults to RDMS or name in USE QUALIFIER

• Version-names not allowed with view-names

• Column names default to names from SELECT list

• Definition of view is stored in repository

• Owner of a view is the user-id that created it; you cannot transfer ownership

• The query-specification cannot have an aggregate operation if you specify a view in

the FROM clause that had an aggregate operation in its query specification;
aggregate operations include:

- DISTINCT keyword in the select list

- GROUP BY clause

HAVING clause

- Built-in functions: SUM, AVG, MAX, MIN, COUNT

SG 3605 13-7
Views

View Example

CREATE VIEW EMP_LOC

(NAME, LOCATION)
AS SELECT ENAME, LOC
FROM EMP, DEPT
WHERE EMP.DNO = DEPT.DNO

Defines the virtual table:

--Name--- -Location-
FOSTER NEW YORK
BROWN NEW YORK
WOODWORTH NEW YORK
ROCKWELL ATLANTA
MARTIN NEW YORK
ADAMS NEW YORK
GLASS ATLANTA
TURNER PRINCETON
LAWSON ATLANTA
JOHNSON PRINCETON
WILLIAMS DETROIT
SMITH PRINCETON
FORD NEW YORK
HYDE PRINCETON

A view can be queried just like a base table.

SELECT DISTINCT LOCATION FROM EMP_LOC

-Location-
PRINCETON
ATLANTA
NEW YORK
DETROIT

13-8 SG 3605
 Views

Columns in Views

• Columns can be renamed

• Number of columns must equal number of items in SELECT list

• Order of column names corresponds to order of items in SELECT list

• Columns can be derived values using:

- Functions

Arithmetic expressions

- Constants

• Column name list is mandatory when SELECT list contains:

Duplicate column names

- Calculations (arithmetic expressions/functions)

- Constants

Example:

CREATE VIEW BONUS

(EMPNO, NAME, SALARY, COMM, BONUS)
AS SELECT EMPNO, ENAME, SAL, COMM, SAL*.10
FROM EMP

SG 3605 13-9
Views

Data Checking

WITH CHECK OPTION

• Used to create domain constraints on column values

• UPDATE and INSERT commands have values checked against WHERE clause of the
definition of the view

Example:

CREATE VIEW CHECKED VIEW

AS SELECT * FROM EMP
WHERE HIREDATE > 900101
WITH CHECK OPTION

Security

WITH ACCESS CONTROL

• Controls who can access the view

• Owner of view is user who creates it

• Owner can use GRANT and REVOKE commands for privileges on view

• Owner inherits privileges from base tables on which view is defined

Example:

CREATE VIEW PRIVATE_VIEW

AS SELECT EMPNO, SAL FROM EMP
WITH ACCESS CONTROL

13-10 SG 3605
 Views

Updating Views

• All changes to a view are reflected in the base table and vice versa

• Views are either read-only or updatable

View is read-only if the CREATE VIEW command includes any of the following

- DISTINCT keyword

- An entry in the select list other than a column specification or *

- A column specification that appears more than once in the select list

- A GROUP BY clause

- A HAVING clause

A FROM clause that references more than one table or view

- A FROM clause that identifies a read-only view

• INSERT and DELETE restrictions

- View must meet all UPDATE conditions

- View must select all columns in base table

SG 3605 13-11
Views

DROP VIEW Command

DROP VIEW [ qualifier. ]view-name

• This command is not available in SQL 89

• Removes a view definition from the repository

• Only owner can drop a view

• Views are dropped automatically when:

- Column in base table is changed or dropped

- Base table or view is dropped

- Owner's access to base table is removed

— Access control is activated for a different owner for an underlying base table

Example:

DROP TABLE QUAL1.TABLE1, TABLE2, QUAL2.TABLE3

13-12 SG 3605
 Views

Exercise 13-1

Views
1. What is a view?

2. List four advantages of using views.

3. What must be true to be able to perform updates on a view?

4. What must be true to be able to perform inserts or deletes on a view?

SG 3605 13-13
Views

Lab 13-1

Use the IPF SQL to perform the following exercise.

1. Create a view based on EMP and DEPT, and another view based on EMP alone.
Incorporate renaming some columns and use of functions and/or arithmetic
expressions.

- Use the IPF File capability to create the views.

- Start your commands with a BEGIN THREAD command and end with an
END THREAD command.

2. Attempt to update the first view.

3. Attempt to delete a row from the first view; from the second view.

4. Create a third view that will be updatable.

5. Perform an INSERT, UPDATE, and DELETE on the third view.

13-14 SG 3605
14
More About Threads
and UDS
 More About Threads and UDS

Module 14
More About Threads and UDS
Module Objectives 14-3

Benefit Statement 14-3

Materials 14-3

Threads 14-4

Threads and Steps 14-5

Thread Control Commands 14-6

Recovery and Thread Control 14-7

BEGIN THREAD Command Revisited 14-8

Application Groups 14-9
Recovery Options 14-10

COMMIT Command 14-11

Message Recovery on COMMIT/END THREAD 14-12

Message Recovery on ROLLBACK 14-13

Locking 14-14
Concurrent Access 14-14
Locking 14-14
Types of Locks 14-15

LOCK Command 14-16

Implicit RDMS Row Locks 14-17

Locking Conflicts 14-18

Deadlock 14-19

Resolving Deadlock 14-20

UNLOCK Command 14-21

Program Flow with Locks 14-22

Exercise 14-1 14-23

SG 3605 14-1
 More About Threads and UDS

Module Objectives

Upon completion of this module, you should be able to

• Select recovery options for recoverable steps.

Describe or use the locking mechanisms:

- Describe the implicit locking by UDS.

- Use the explicit locking commands.

Benefit Statement

Two large issues related to database integrity center around disruption of the computer
itself and multiple users accessing the data. In this module you will realize the choices
you have at your disposal in regard to these issues, as well as the automatic coverage
provided by UDS Control and the Integrated Recovery Utility.

Materials

• OS 1100 UDS Relational Data Management System (UDS RDMS 1100) and IPF SQL
Interface End Use Guide

• OS 1100 UDS Relational Data Management System (UDS RDMS 1100) SQL
Programming Reference Manual

• OS 1100 UDS Relational Data Management System (UDS RDMS 1100)

Administration Guide

• OS 1100 Universal Compiling System (UCS) Application Development Programming

Guide, Volume 1

SG 3605 14-3
More About Threads and UDS

Threads

• A thread is a work session with UDS Control

- Each thread belongs to one user

- Each thread exists in one application group

- Several threads may be active at one time

• UDS Control establishes your database session environment

Thread control

- Application groups

- Error message handling

Recovery options

- Concurrency control

• Thread control commands are processed directly by UDS Control, without invoking
RDMS

14-4 SG 3605
 More About Threads and UDS

Threads and Steps

step step step step step

>< >< >< ><

< 0 0 0 0 >

Start BEGIN COMMIT COMMIT COMMIT ROLLBACK END End

Program THREAD THREAD Program

Figure 14-1. Step - Thread Control Relationships

• A thread's session with UDS 1100 can be subdivided into steps

• A step is logical unit of work

- Sequence of commands that belong together

- All changes are made permanent or

- All changes are undone

Smallest recoverable unit of a thread

• Integrated Recovery Utility (IRU) recovers to the previous step

• Each step begins and ends with a thread control command

SG 3605 14-5
More About Threads and UDS

Thread Control Commands

BEGIN THREAD

— Establishes a new session with UDS

• COMMIT WORK

— Ends a step

— Updates since last thread control command are written to database

• ROLLBACK

— Ends a step

- Updates since last thread control command are discarded

• END THREAD

- Ends thread and commits changes

- Program must BEGIN THREAD again before issuing more SQL commands

14-6 SG 3605
 More About Threads and UDS

Recovery and Thread Control

• Recovery is the process of returning the database to a consistent state

- Follows a user error or a hardware/software failure

- Recovery is a function of UDS Control

- Integrated Recovery recovers to beginning of step

• The type of recovery performed depends on:

— Creating table storage-area as recoverable

— BEGIN THREAD command recovery-option

• As a programmer, it is important to know if the file is recoverable

• BEGIN THREAD command declares the type of recovery to use

If thread is recoverable, ROLLBACK and COMMIT commands may break it into

smaller units of processing (steps)

• At end of ROLLBACK or COMMIT, the data is in a consistent state

— All changes are discarded (ROLLBACK)

— All changes are written (COMMIT)

SG 3605 14-7
More About Threads and UDS

BEGIN THREAD Command Revisited

BEGIN THREAD [ thread-name ]

FOR [ APPLICATION ] application-name [ options ]

• Unisys implementation

• Where application-name may be an alias or application group to use

- Application names are limited to 6 characters

- Alias names are limited to 12 characters

• The UPDATE option has the format:

UPDATE ( recovery-option ) [ UDSMSG UDSMESSAGE UDSMESSAGES ]

• Where recovery-option may be:

COMMANDLOOKS

- DEFERRED

- QUICKLOOKS

- NONE

Example

BEGIN THREAD t1 FOR APPSVN UPDATE (DEFERRED)

14-8 SG 3605
 More About Threads and UDS

Application Groups

• An OS 1100 software partition

- Purpose is to create multiple working environments

- Each environment acts as an independent, recoverable system

- Up to nine application groups supported on a system

- Each has its own common banks, database files, audit files, repository, and other
support components

- Default application group name is UDSSRC

• BEGIN THREAD determines application group for rest of program execution

• ACOB program with SQL must be collected with proper application group elements

- Two elements are CBEP$$RSA and RDMR-ACOBDAT

- Typical default application group file name is SYS$LIB$*RSA

- DBA must be consulted for filename for other application groups

• UCOB program with SQL must be linked to proper application group elements

- Linking to the UDSSRC requires nothing by user

- ESQL program uses keyword option APPLICATION/application-name at compile

time for, alternate application group (not UDSSRC)

- ESQL or ISQL dynamically links to alternate application groups with proper

@USE before execution

Example: Compile and execute a UCOB ESQL program that will use application group
seven

@UCOB,se sql.prog 1,tpf$.prog 1„,APPLICATION/AP PSVN

@USE LINK$PF,SYS$LIB$*APP$7.
@XQT tpf$.progl

SG 3605 14-9
More About Threads and UDS

Recovery Options

• QUICKLOOKS

- Snapshot of page is taken before first update of page in a step

- Database is recoverable to the before-image taken

- Recommended for large number of updates

• DEFERRED

- No updates written to database until COMMIT or END THREAD issued

- Changes saved in temporary file

- Database has only most recent updated copy of each page

- Recommended for few updates or intentional ROLLBACKS

Default recovery option (New with RDMS 5R1)

• COIVIMANDLOOKS

- Snapshot of page is taken before every update command

- The first before-image is saved until the end of the step

- If an update command succeeds, the previous commandlook is discarded

- If update fails, database is restored to before-image state

- Similar to Quicklooks, but more overhead

- Recommended for interactive programs where SQL command depends on user

input

• NONE

- No recovery is possible, not a recommended option

- Use only on storage-areas defined as non-recoverable

14-10 SG 3605
 More About Threads and UDS

COMMIT Command

COMMIT WORK

• Writes all updates to database since beginning of step

• Thread remains registered with UDS

• ROLLBACK can only recover to previous COMMIT (end of last step)

• Releases all implicit locks set by RDMS

Example:

COMMIT WORK

SG 3605 14-11
More About Threads and UDS

Message Recovery on COMMIT/END THREAD

COMMIT [ WORK I THREAD ] [ step-control-option ]

END THREAD [ step-control-option ]

SQL extension for COMMIT

• Where step-control-option may be:

— ADVANCE

— TERMINATE

• Message options apply to Integrated Recovery only

• ADVANCE

— Terminates current step

— Begins new step

— Input message retained

• TERMINATE

- Terminates current step

- Releases input message to Message Control Bank (MCB)

14-12 SG 3605
 More About Threads and UDS

Message Recovery on ROLLBACK

ROLLBACK [ WORK ] [ step-control-option ]

• SQL extension

Where the step-control-option specifies what to do with the message in Integrated

Recovery environment

Note: Unless you are using TIP with MCB, it may not be relevant to specify this
option.

• Not specified

— Use the default step-control option

• KEEP

— Resumes current step

— Input message is still associated with active step

• DISCARD

- Terminates current step

- Releases any deferred messages

- If input message is recoverable, queue it to a system error program

- If input message is not recoverable, release it

• RE QUEUE

- Terminates current step

- Message must be recoverable, so requeue for reprocessing of same step

SG 3605 14-13
More About Threads and UDS

Locking

Concurrent Access
In a normal database environment, many users will access the same table, maybe even
the same rows, simultaneously. If everyone is just looking at the data, there is no
problem with concurrent access. The problem occurs when two users want to update, or
one user wants to update and another look at the same information. In this case, the
database management system software must control who can read and who can update
the data.

To ensure that no user is looking at information that is currently being changed, RDMS
1100 places contention locks on the data. Two rules apply:

1. At any given time, only one user can update a row.

2. While a user is updating a row, no one else may read it.

Locking
UDS Control and RDMS 1100 provide a system of locking that gives basic protections to
users. You can let RDMS automatically do the locking for you. This is implicit locking.
For more efficient processing and specific locking, you can incorporate LOCK commands
explicitly in a program.

14-14 SG 3605
 More About Threads and UDS

Types of Locks

• Specified by a combination of mode and usage

- The mode establishes how strong the lock is and is optional

- The usage establishes whether you are retrieving or updating data

Allowed for mode: EXCLUSIVE, PROTECTED, SHARED, not-specified

- Allowed for usage: RETRIEVAL, UPDATE

• RETRIEVAL (R)

— Thread will only read data

• UPDATE (U)

— Thread may alter data or table definitions

• SHARED (S)

- Other threads may concurrently access the table

- Row locks kept even when row is no longer current

- Prevents others from viewing updated but not yet committed data

• PROTECTED (P)

— No other thread may alter the table while the lock is on

• EXCLUSIVE (X)

— No other thread can access the table while the lock is on

- Strongest lock

• Eight types of locks are allowed by RDMS: R, SR, U, SU, PR, PU, XR, and XU

SG 3605 14-15
More About Threads and UDS

LOCK Command

LOCK [ TABLE ] table-specification-list

[ IN lock-specification [ MODE ] ]
[ ON CONFLICT conflict-option ]

• Unisys implementation

• Where table-specification-list may specify a list of tables separated by commas

- A view cannot be locked

- A separate LOCK command should be used for each table

- An error could occur for the command

- Some tables could be locked or others not, or queuing could take place for
some but not all tables

• The lock-specification is in the format [ mode ] [ usage ], default RETRIEVAL

• The conflict-option specifies how to handle conflicts, default QUEUE

— RETURN means RDMS issues an error and control returns to your program
- , ,

— QUEUE means I MS queues your program until the locking conflict is resolved

• Usage guidelines

— LOCK command precedes OPEN command execution

— Lock remains until table is explicitly unlocked or thread is ended

Examples:

LOCK EMP IN UPDATE MODE

LOCK EMP IN EXCLUSIVE RETRIEVAL MODE ON CONFLICT RETURN

14-16 SG 3605
 More About Threads and UDS

Implicit RDMS Row Locks

RDMS implicitly locks rows of a table as they are read or updated. The type and duration
of implicit lock depends on the type of explicit table-level lock used. Table 14-1 shows the
implicit locks set by RDMS.

Table 14-1. Implicit Row-Level Locks Set by RDMS

RDMS Sets This Implicit RDMS Sets This

Table Lock Row-level Lock on Implicit Row-level
Specified Retrieved Rows Lock on Updated Rows

No lock specified Command-duration Step-duration exclusive

retrieval lock on each update lock on each
retrieved row * updated row
Retrieval Command-duration Updated not allowed
retrieval lock on each
retrieved row *
Shared update Step-duration protected Step-duration exclusive
update lock on each update lock on each
retrieved row updated row
Update Command-duration Step-duration exclusive
retrieval lock on each update lock on each
retrieved row * updated row
Shared retrieval Step-duration retrieval lock Updates not allowed
on each retrieved row
Protected retrieval No lock on retrieved rows Updates not allowed
Protected update No lock on retrieved rows Step-duration exclusive
update lock on each
updated row
Exclusive retrieval No lock on retrieved rows Updates not allowed
Exclusive update No lock on retrieved rows No locks on updated rows

For a cursor declared with the FOR UPDATE clause, RDMS places a retrieval
lock on each row as it is fetched, releasing the lock only when you alter the
currency of the cursor by fetching a different row from the cursor, closing the
cursor, dropping the cursor, or ending the step.

SG 3605 14-17
 More About Threads and UDS

Locking Conflicts

A locking conflict occurs if two users try to lock the same table at the same time. Table
14-2 summarizes the conditions in which locks from different runs conflict. An * indicates
a conflict.

Conflicts apply to both table and row level locks, but table-level locks can only conflict
with table-level locks, and row-level locks with row-level locks

Table 14-2. Locking Conflicts

Current Lock -3 R SR U SU PR PU XR XU
New Usage i

Retrieval * *

Shared Retrieval * *

Update * * * *

Shared Update * * * *

Protected Retrieval * * * * *

Protected Update * * * * * *

Exclusive Retrieval * * * * * * * *

Exclusive Update * * * * * * * *

• Horizontal axis represents the type of lock currently in place

• Vertical axis represents the type of lock requested by another run

L D cr-

14-18 SG 3605
 More About Threads and UDS

Deadlock

Deadlock occurs when two or more runs are queued on locks held by each other so that
none of the runs can resume execution.

Table 1

Table 2

---).Lock held
Lock attempted

Figure 14-2. Illustrating Deadlock

SG 3605 14-19
More About Threads and UDS

Resolving Deadlock

• Reduce probability of deadlock by using "strong", explicit locks

• Tradeoff - Strong locks reduce concurrent access

• RDMS returns deadlock status

— ERR-STAT = 6010

— Test in program

• Suggested resolution

— Wait from 1/10 to 1 second

Repeat all OPEN, INSERT, DELETE, and UPDATE commands in step rolled
back

- Include in loop

- Use randomly generated wait times if several threads are doing this
'-;„ • " • -
• On conflict, the RETURN option of the LOCK command returns:

— ERR-STAT = 6004 when locking conflict occurs . - .

— RDMS performs END THREAD/ROLLBACK

14-20 SG 3605
 More About Threads and UDS

UNLOCK Command

UNLOCK table-specification-list

• Unisys implementation

• Removes lock from table(s)

• Retrieval locks are dropped immediately

• Update locks are dropped at end of step

— COMMIT

- END THREAD

• If table is being used by an active cursor, UNLOCK generates an error

• UNLOCK is required to change lock mode on a table

Example:

UNLOCK EMP, DEPT

SG 3605 14-21
More About Threads and UDS

Program Flow with Locks

BEGIN THREAD

DECLARE CURSOR

LOCK

DROP CURSOR

V
UNLOCK

END THREAD

Figure 14-3. Program Flow with Locks

14-22 SG 3605
 More About Threads and UDS

Exercise 14-1

1. What is a step? What is the purpose of breaking a program into steps?

2. What occurs when a step is committed?

3. What occurs when a step is rolled back?

!-•

4. Why does RDMS lock tables?

SG 3605 14-23
15
Getting Physical -
Tables and Storage
Areas
 Getting Physical - Tables and Storage Areas

Module 15
Getting Physical - Tables and Storage Areas

Module Objectives 15-3

Benefit Statement 15-3

Materials 15-3

Creating a Database 15-4

Schema 15-5

Storage-Areas 15-6

Storage-Areas, Files, and Tables 15-7

Logical Tables and Physical Files 15-8

Table Versions 15-9

Fully Qualified Table-Name 15-10

USE DEFAULT Command 15-11

How RDMS 1100 Locates Data 15-12

How RDMS 1100 Resolves Table References 15-13

Creating Versions Example 15-14

Exercise 15-1 15-15

SG 3605 15-1
 Getting Physical - Tables and Storage Areas

Module Objectives

Upon completion of this module, you should be able to

• Describe the process of creating tables

- Identify the components of a fully qualified table name
Set a default table version using the SET DEFAULT VERSION command

— Describe the difference between a logical and physical table

- Describe the relationship between physical tables and storage areas

Benefit Statement

Access and manipulating tables have been the main topic for this course, but what if you
are responsible for setting up the database. A few insights into creating the tables and
how they relate to the database physical attributes are presented here. Your next step
may be database administration!

Materials

• OS 1100 UDS Relational Data Management System (UDS RDMS 1100)

Administration Guide

• OS 1100 UDS Relational Data Management System (UDS RDMS 1100) SQL
Programming Reference Manual

SG 3605 15-3
Getting Physical - Tables and Storage Areas

Creating a Database

Catalog files for ECL

Storage - Area(s)

Create Schema UREP

Description

Create Storage - UREP

Area Description

Add Storage - Area UREP

Attributes

Process Storage - Area UREP

Check/Install

Create tables SQL

Load data SQL

into tables

Figure 15-1. Steps to Creating a Database

15-4 SG 3605
 Getting Physical - Tables and Storage Areas

Schema

• Definition
Database definition that describes units of data, such as record types and
relations
A name that identifies an arbitrary group of storage area definitions and table
definitions in repository

Schema

Storage-Area Table
Descriptions Descriptions

Figure 15-2. Schema Function

SG 3605 15-5
Getting Physical - Tables and Storage Areas

Storage-Areas

• Storage-Areas are defined by the Database Administrator

• Definition contains:

Storage-Area Name

- Schema with which it is associated

Exec file name or TIP file number

— Version-name

• Each storage-area is associated with exactly one file

- Name of storage-area and file may be different

- Version-name of storage-area is same as that of table which will be stored in it

• Storage-area must be defined before corresponding table

• As programmer, need to know version-name for reference

15-6 SG 3605
 Getting Physical - Tables and Storage Areas

Storage-Areas, Files, and Tables

Create
Storage-Area
Company.Emp:Prod
Company.Emp:Tpst

Associate
with File

PROD • EMP .41(

TEST " EMP

Create Table
and Enter Data

Company.Emp:Prod

Compeny.Emp:Test

Figure 15-3. Connecting the Pieces

SG 3605 15-7
Getting Physical - Tables and Storage Areas

Logical Tables and Physical Files

• Logical table defines:

Column layout
- Datatypes

- Primary key(s)
- Secondary key(s)
- Foreign key(s)

• Versions permit spreading data across multiple physical files

• File-name is specified by storage-area version

• Characteristics of each file may be different

Audited/Not audited
Recoverable/Not recoverable

Pages allocated

15-8 SG 3605
 Getting Physical - Tables and Storage Areas

Table Versions

• Versions of a table use the same table definition

• Versions of the table are created when different data is loaded in the specified table
versions

• Versions can be used to divide a large table into tables with fewer rows

EMP : EMP : EI4P :

DIVA DIVE DIVC

Figure 15-4. Table Versions

SG 3605 15-9
Getting Physical - Tables and Storage Areas

Fully Qualified Table-Name

schema-name. table-name:

• The schema-name is used as qualifier for table name

• Default qualifier is RDMS

• Default version

For RDMS is PRODUCTION

- For UREP is blank

These two are equivalent

• Table qualifier has nothing to do with EXEC file qualifiers

• Fully qualified table name may be used anywhere table name must be coded

• Can set up own defaults for qualifier/version so only table name is needed

Examples:

Company. Payroll:Production

Company.Payroll:Test

15-10 SG 3605
 Getting Physical - Tables and Storage Areas

USE DEFAULT Command

USE DEFAULT[ QUALIFIERuse-name I VERSION version-name]

• Unisys implementation

The use-name may be:

- Schema-name

- Variable containing string to use of 30 or fewer characters

• Remains in effect until next USE command

• The version-name may be

— Name or string

— Variable containing version-name

Generalized programs

- Accept input of qualifier-name or version-name into program variables

- Then USE DEFAULT to access desired table

Examples:

USE DEFAULT QUALIFIER SALES_DEPT

USE DEFAULT QUALIFIER 'R & D'

USE DEFAULT VERSION TEST

USE DEFAULT VERSION $P1

. SG 3605 15-11
Getting Physical - Tables and Storage Areas

How RDMS 1100 Locates Data

Table name you give to RDMS:

Payroll

RDMS assumes this jr 1

fully qualified name

RDMS.PAYROLL•PRODUCTION

Items in RDT 2
vir
Table name
Qualifier
Storage-Area name
Column names
Primary key
Foreign key
Indexes

Items in FDT
V
Storage-Area name
Schema name
Version name
File name
Other attributes

Mass-storage file IL 4

Data associated
with this
storage-area
c_

Figure 15-5. RDMS Locating Data

15-12 SG 3605
 Getting Physical - Tables and Storage Areas

How RDMS 1100 Resolves Table References

1. When you give RDMS a table name, it must first determine the full form of the name in
the format schema.table:version.
If you do not specify the schema or version, RDMS uses the current default values.

Unless a USE command was entered, the default schema is RDMS, and the default
version is PRODUCTION.

2. RDMS uses the schema and table names together to find the encoded form of the table
description (Relational Table Description or RDT) in the RDT$FILE.

3. RDMS gets the storage-area name from the RDT, and combines it with the version name
to identify the File Description Table (FDT) in FDT$FILE.
The FDT is the encoded form of the Storage-Area Definition.

This contains the name of the file in which the table data is physically stored.

4. RDMS gets the EXEC file name or TIP file number from the FDT. It then goes to the
file and locates the table.

SG 3605 15-13
Getting Physical - Tables and Storage Areas

Creating Versions Example

@CAT,P KOP*INSTRUCTORS.
@OAT,P NYC*INSTRUCTORS.

First use UREP

@DD,E
CREATE SCHEMA MASTER INSTRUCTOR.
CREATE STORAGE-AREA INSTRUCTOR
VERSION KING_ OF PRUSSIA
FOR SCHEMA MASTER_ INSTRUCTOR.
ADD QUALIFIER KOP. -
ADD FILENAME INSTRUCTORS.
ADD FILE-TYPE EXEC.
PROCESS STORAGE-AREA INSTRUCTOR
VERSION KING_ OF PRUSSIA
FOR SCHEMA MASTER INSTRUCTOR INSTALL.
CREATE STORAGE-AREA INSTRUCTOR
VERSION NEW YORK CITY
FOR SCHEMA MASTER_ INSTRUCTOR.
ADD QUALIFIER NYC.
ADD FILENAME INSTRUCTORS.
ADD FILE-TYPE EXEC.
PROCESS STORAGE-AREA INSTRUCTOR
VERSION NEW YORK CITY
FOR SCHEMA E
RAST INSTRUCTOR INSTALL.
EXIT.

Then, in SQL perform the following:

CREATE TABLE MASTER INSTRUCTOR.INST INFO

IN MASTER INSTRUCTOR.INSTRUCTOR.
COLUMNS ARE INST NAME
•
•

USE DEFAULT QUALIFIER MASTER INSTRUCTOR;

INSERT INTO INST_ INFO:KING_OFPRUSSIA
VALUES('SANDY KNORR', ...);
INSERT INTO INST_ INFO:NEW_YORKCITY
VALUES('DONNA KEECH', ...);

15-14 SG 3605
 Getting Physical - Tables and Storage Areas

Exercise 15-1

1. What is a schema? How is it used by RDMS?

2. List three differences between a logical and a physical table.

(7"

3. Write the commands needed to make the table-name EMP resolve to US.EMP:DIV1.

(`j

SG 3605 15-15
A
System Usage
Information
 System Usage Information

Appendix A
System Usage Information

System Essentials A-3

Demand Session and Instructions A-5

MAPPER Session and Instructions A-6

Obtaining Printout A-8

Compiling and Executing a Program in IPF A-9

Compiling and Executing a Program using ECL A-10

Use Runstreams A-10
ACOB program A-10
UCOB program A-10

Saving and Printing Compiler Listings A-11

Basic IPF Commands A-12

Command Line Commands A-12
Text Line Commands A-12

SG 3605 A-1
 System Usage Information

System Essentials
Your instructor will assist you in filling out the information needed for this class.

• Demand Login

— $$open-ID

— User-ID/Password

— Run-ID

— Account

— Project-ID

• MAPPER Login

$$OPEN-ID

— MAPPER-ID

- User-ID

Password

— Department-#

• Printer information

Printer-ID

Printer location

Other

• Database information

- Qualifier/Schema-name

- Table Names

- Application group

SG 3605 A-3
System Usage Information

Demand Session and Instructions

You may need to note changes required by your site for the following "typical" session.

Note: The greater than sign ">" is used instead of the standard start of entry (SOE)
character below.

>$$OPEN DMEA10
>USER-ID/PASSWORD
>Project-ID: STU . System either solicits project... or
>@RUN run-ID,account,stu ... or requires run statement

Note: You will need a couple of files for this class, a program file for your programs
and a data file for your printout. The suggested names of the files will both
have your STU number as the qualifier. The program file should be named
SQL and the printfile named PRFIL. These are the names you will find in the
provided runstreams. Change them if you wish, but don't step on someone
else's files!

>@CAT,P SQL. . Use your STU## as the default qualifier

>@CAT,P PRFIL. . Catalog your files once and assign them at the
>@ASG,A SQL. beginning of each demand session
>@ASG,A PRFIL.
>@ASG,A AL3644*SQL. Contains program skeletons and addstreams
>@COPY,P AL3644*SQL.,STU _ *SQL.
>@PRT,T SQL.

Note: The above command lists the elements provided for this class. When an
element name has two parts divided by the "/", the first part tells you where
this element is used and the second part tells you what it is used for. For
example, the element LAB3-1 I UCOBSKEL is the element for Lab 3-1 and is a
UCS COBOL skeleton. Other elements will be described by your instructor.

A-4 SG 3605
 System Usage Information

Demand session stuff: @IPF, Compile, Execute, etc.

>
>@FIN
>$$CLOSE . Please close the session when done!

Additional Demand Notes:

MAPPER Session and Instructions

You may need to note changes required by your site for the following "typical" session.

>$$OPEN T4EA10
>EDMAP . Select a specific MAPPER
Enter userid,dept#,password on Menu or on the control line
after ")"
>RDI . Entered from control line

. Other MAPPER session commands

>x . Entered from control line

▪ Exit from MAPPER

>$$CLOSE . Close the session

Additional MAPPER Notes

SG 3605 A-5
System Usage Information

Obtaining Printout

Note: This information is site dependent. You may have an on-site printer, a local
attached printer, a remote printer, or no printer available! And how you access
the printer will require specific instructions. Please make your notes below.

Some of the commands that obtain printout include:

>@SYM, U PRFIL„printer-id„print-banner . Queue to onsite printer

>@add,l sql.spool DDP and Queue the file to a
remote printer
>-C->site printer-ID IPF command for on-site printer
>aux MAPPER uses local printer
attached to terminal. Menu .
>pr MAPPER queues output to site
printer

Printer Notes

A-6 SG 3605
 System Usage Information

Compiling and Executing a Program in IPF

• Login to IPF
>@IPF filename

• Retrieve program file into work space

>-C-->old Qual*filename.elt

• Set datamanager
>-C->Set $data := rdms

• Run program
›-C->run filename.elt, compiler, library-list
>-C-->run sql.lab3-1/ucobskel, ucob

IPF automatically compiles, Jinks, and executes element

— Defaults

— Executes work space

— Compiler corresponds to file type

— System library
>-C-->run ,ucob

Compiles and executes work space using ucob compiler and system library

• Print the program in the work space

>-C->Site printer-ID

SG 3605 A-7
System Usage Information

Compiling and Executing a Program using ECL

Use Runstreams

Note: Runstreams are provided in your student file to automate this process. Please
modify them for your use during this class. If the version name of the element
contains "r-u-n" it is likely to be a runstream. Example of using one:

>@ADD,L sql.lab3-1/runucobskel

ACOB program
The following pieces are intended to operate together in a demand session.

• Compile source program

>@ACOB,ES sql.lab3-1/acobskel,rel Relocatable output to TPF$

• Collect the program

>@ADD sql.map . Element map is in your work file

• Execute absolute

>@xqt abs . MAP output was to TPF$.abs

UCOB program
• Compile source program

>@UCOB,res sql.lab3-1/ucobskel,om . Output to TPF$.om

• Execute object module

>@xiqt. om . Invokes dynamic linking to

libraries

A-8 SG 3605
 System Usage Information

Saving and Printing Compiler Listings

Using and modifying the provided runstreams automates the process of using the
standard ECL. You can save your printfile information in your program file with an ECL
copy:
>@copy,i prfil.,sql.lab3-1/ranucobskel . Note the "ran" versus
"run" name

A popular, but unsupported, pair of processors can be found at most sites: @BK1 and
@BK2.

• Execute a breakpoint: @BK1

• Perform interim commands

• End breakpoint and view output in editor: @BK2,opts

- E for ED, C for CTS, I for IPF

• End breakpoint and queue output to printer: @BK2 printer-ID

Example 1
>@BK1
>@UCOB,res sql.lab3-1/ucobskel,om
>@BK2,I

Example 2

>@BK1
>@ACOB,se sql.lab3-1/acobskel,rel
>@EOF
>@ADD sql.map
>@BK2 prinl

SG 3605 A-9
System Usage Information

Basic IPF Commands

Command Line Commands

new Start a new file
old Retrieve text of filename into work space
save Save contents of work space (a new file)
repl Save contents of work space (existing file)
top Go to the top of the file
bottom Go to the bottom of the file
go Go to the given line number

Text Line Commands

In Insert n lines after this line
IBn Insert n lines before this line
G-i-n Go to n lines past this line
G-n Go to n lines before this line
Dn Delete n lines
DS Delete start - Start of text to delete
DE Delete end - End of text to delete
Cn Copy n lines
CS Copy start - Start of text to copy
CE Copy end - End of text to copy
CB Copy the marked text before this line
CA Copy the marked text after this line
Mn Move n lines
MS Move start - Start of text to move
ME Move end - End of text to move
MB Move the marked text before this line
MA Move the marked text after this line

A-10 SG 3605
B
FORTRAN Examples
 FORTRAN Examples

Appendix B
FORTRAN Examples

SQL Commands in ASCII FORTRAN B-3

Sample ASCII FORTRAN Retrieval Program B-4

Example With GETERROR (FTN) B-6

SG 3605 B-1
 FORTRAN Examples

SQL Commands in ASCII FORTRAN

Format 1

CALL F$RDMR
('SQL-COMMAND' ,
ERSTAT ,
AUXINF )

Format 2

CALL F$RDMR
(SQL-COM-STRING-VARIABLE ,
ERSTAT
AUXINF )

The SQL command or the contents of SQL-COM-STRING-VARIABLE must end with a

semicolon (;).

• ERROR-STATUS - A CHARACTER*4 item that contains a status code when RSA

detects an error. Normal completion has status 0000.

• AUXILIARY-INFORMATION - An INTEGER item that contains the column position

within the source command where RSA detects an error. If there is no error, the
contents are meaningless.

Example

CHARACTER*132 RCOM
CHARACTER*4 ERSTAT
INTEGER AUXINF
CALL F$RDMR
* ( 'BEGIN THREAD FOR UDSSRC READ
ERSTAT, AUXINF )

RCOM = 'DECLARE EMPLOYEE CURSOR SELECT * FROM EMP;'

CALL F$RDMR ( RCOM, ERSTAT, AUXINF )

SG 3605 B-3
FORTRAN Examples

Sample ASCII FORTRAN Retrieval Program

If FORTRAN is your programming language, here is a complete program in ASCII

FORTRAN.

(At this stage, you don't need to be concerned about the syntax and layout of the
commands. This is explained in the modules. For now, look at the way the RDMS
commands are embedded in the 'CALL F$R.DMR...' commands.)

Requirement: Read and display all the rows of the EMP table, showing all the columns

C
c****************************************************************
C * MAIN PROGRAM TO PRINT CONTENTS OF THE "SALESMEN" TABLE
c****************************************************************
CHARACTER*511 rcom
CHARACTER*4 erstat
INTEGER auxinf
INTEGER rempno
CHARACTER*10 rename
CHARACTER*10 rjob
INTEGER rmgr
INTEGER rdno
INTEGER rhirdt
INTEGER rsal
INTEGER rcomm
C
c****************************************************************
C * CONNECT TO UDS
*****************************************************************
RCOM = 'BEGIN THREAD FOR UDSSRC READ UDSMSG ;'
CALL F$RDMR(RCOM, ERSTAT, AUXINF)
CALL errchk(erstat, auxinf, ", *90, *90)
C
C
c****************************************************************
C DECLARE THE CURSOR
c****************************************************************
PRINT '('Retrieves entire relation EMP', ///)'
RCOM = 'DECLARE EMPLOYEE CURSOR SELECT * FROM EMP;'
CALL F$RDMR(RCOM, ERSTAT, AUXINF)
CALL errchk(erstat, auxinf, ' ', *90, *90)

B-4 SG 3605
 FORTRAN Examples

C
C****************************************************************
C LOOP TO READ TABLE AND PRINT ROW *
C****************************************************************
10 CONTINUE
RCOM = 'FETCH NEXT EMPLOYEE INTO $P1, $P2, $P3, $P4, $P5,
* $P6,$P7,$P8 ;'
CALL F$RDMR(RCOM, ERSTAT, AUXINF,
* REMPNO, RENAME, RJOB, RMGR, RDNO, RHIRDT, RSAL, RCOMM)
CALL errchk(erstat, auxinf, '6001', *90, *80)
C
PRINT ' (1X, I4,1X,A, I3,1X, I6, 1X, I4,1X, I4) ' ,
* REMPNO, RENAME, RJOB, RMGR, RDNO, RHIRDT, RSAL, RCOMM)
GO TO 10
C
C
80 CONTINUE
PRINT 'End of relation EMP.'
CALL F$RDMR('END THREAD;', ERSTAT, AUXINF)
CALL errchk(erstat, auxinf, ", *90, *90)
PRINT '(23HONormal end of program.)'
STOP
C
90 CONTINUE
PRINT '(40H0***** Error status program termination.)'
CALL F$RDMR('END THREAD;', ERSTAT, AUXINF)
STOP
END
C
C*
C SUBROUTINE TO TEST ERROR STATUS
C
C
SUBROUTINE errchk (erstat, auxinf, accept, *, *)
C
CHARACTER*4 accept @ Acceptable status for RETURN 2
INTEGER auxinf
CHARACTER*4 erstat
C
IF (erstat .EQ. '0000') RETURN
IF (erstat .EQ. accept) RETURN 2
C
PRINT '(29H0**** Unexpected error status ,A,11H
at column,I3)', erstat, auxinf
RETURN 1

SG 3605 B-5
FORTRAN Examples

Example With GETERROR (FTN)

CHARACTER*132 El, E2, E3, E4, E5

ROOM = 'BEGIN THREAD T1 READ ;'

CALL F$RDMR(RCOM, ERSTAT, AUXINF)
IF ERRCOD .EQ. '0000' THEN (SKIP ROUND ERROR ROUTINE)

CALL F$RDMR ('GETERROR INTO $P1, $P2, $P3, $P4, $P5 ;

* ERSTAT, AUXINF, El, E2, E3, E4, E5)

PRINT "(A132)', El
PRINT '(A132)', E2
PRINT '(A132)', E3
PRINT '(A132)', E4
PRINT '(A132)', E5

B-6 SG 3605
C
ACOB Skeleton
 ACOB Skeleton

@PRT,S AL3644*SQL.LAB3-1/RUNACOB
FURPUR 30R2A (910822 1304:40) 1993 Mar 17 Wed 1510:31
AL3644*SQL(1).LAB3-1/RUNACOB(0)
•1 @BRNPT PRINT$,prfil.
2 @PRT,S sql.lab3-1/runacob
3 @ACOB,SE sql.lab3-1/acobskel,TPF$.acobakM1
4 @MAP,S ,TPF$ABS
5 NOT TPF$.
6 IN TPF$.acobskel
7 IN SYS$LIB$*RSA.CBEP$$RSA
8 IN SYS$LIBPRSA.RDMR-ACOBBIAT
9 END
10 @XQT
11 @NWT PRINT$
12 @IPF TPF$.
13 OLD prfil.
14 L END,A,A,NUM
15 MOD SCR
@PRT,S AL3644*SQL.LAB3-1/ACOBSKEL
AL3644*SQL(1).LAB3-1/ACORSKEL(0)
1 IDENTIFICATION DIVISION.
2 PROGRAM-ID. acobskel.
3 ENVIRONMENT DIVISION.
4 CONFIGURATION SECTION.
5 SOURCE-COMPUTER. UNISYS-2200.
6 OBJECT-COMPUTER. UNISYS-2200.
7 *
8 **********************************************************
9 *** This program uses interpretive SQL and ACOB to
10 *** solve the lab3-1 assignment for SQL Programming
11 **********************************************************
12 *
13 DATA DIVISION.
14 WORKING-STORAGE SECTION.
15 *
16 **********************************************************
17 *** Establish RDMS command error buffers and data
18 **********************************************************
19 *
20 *** Error handling variables and message buffers
21 *
22 01 room.
23 05 rlin PIC X(40) OCCURS 12 TIMES.
24 01 err-stat PIC 9(4).
25 01 aux-info PIC S1(36).
26 01 error-text.
27 05 err-s PIC X(132) OCCURS 4 TIMES.
28 01 err-s-index PIC 9(4).
29 01 thread-flag PIC 9 VALDE ZERO.
30 88 thread-has-begun VALUE 1.
31 01 cursor-flag PIC 9 VALUE ZERO.
32 88 end-of-cursor VALUE 1.
33 01 err-end-flag PIC 9 VALUE ZERO.
34 88 no-more-mage VALUE 1.
35 *
36 *** Place for the retrieved record, set up for printing
37 *
38 01 record-retrieved.
39 05
40 05

SG 3605 C-1
ACOB Skeleton

41 ************************************************************
42 *** BEGIN the thread, do the USE, clear the SQL buffer
43 ************************************************************
44
45 *** Start the thread
46
47 PROCEDURE DIVISION.
48 000-main.
49 MOVE ' ;' TO room.
50 ENTER MASK 'acob$rdmr' USING room, err-stat, aux-info.
51 IF err-stat NOT = 0 GO TO 900-error-print.
52
53 *** Set table qualifier, clear buffer, set thread flag
54
55 MOVE ' ;' TO room.
56 ENTER MASK 'acob$rdmr' USING room, err-stat, aux-info.
57 GO TO 900-error-print.
58
59 MOVE 1 TO thread-flag.
60
61 ***********************************************************
62 *** Access RDMS - DECLARE and OPEN the cursor
63 ***********************************************************
64
65 *** Declare the cursor
66
67 MOVE ' ' TO room.
68 MOVE ' ' TO rlin(2).
69 MOVE ' ;' TO rlin(3).
70 ENTER MASK 'acob$rdmr' USING room, err-stat, aux-info.
71 IF err-stat NOT = 0 GO TO 900-error-print.
72
73 *** Open the cursor
74
75 MOVE ' ;' TO room.
76 ENTER MASK USING room, err-stat, aux-info.
77 IF err-stat NOT = 0 GO TO 900-error-print.
78
79 **********************************************************
80 ** Loop thru the rows, end the thread, then do cleanup
81 **********************************************************
82
83 *** Loop to fetch a row and display it
84
85 200-loop.
86 MOVE ' ' TO room.
87 MOVE ' ;' TO rlin(2).
88 ENTER MASK 'acob$rdmr' USING room, err-stat, aux-info
89
90 IF err-stat = 6001 GO TO 400-drop-cursor.
91 IF err-stat NOT = 0 GO TO 900-error-print.
92
93 DISPLAY record-retrieved UPON PRINTER.
94 GO TO 200-loop.
95
96 400-drop-cursor.
97 DISPLAY 'No more data on fetch' UPON PRINTER.
98 MOVE 'DROP CURSOR employee •' TO room.
99 ENTER MASK 'acob$rdmr' USING room, err-stat, aux-info.
100 IF err-stat NOT = 0 GO TO 900-error-print.
101
102 MOVE SPACES TO room.
103

C-2 SG 3605
 ACOB Skeleton

104 ************************************************************
105 *** End the thread, shut down the program
106 ************************************************************
107 *
108 *** Issue the end of the thread
109 *
110 MOVE ' :' TO room.
111 ENTER MASM 'acob$rdmr' USING room, err-stat, aux-info.
112 IF err-stat NOT = 0 GO TO 900-error-print.
113 *
114 *** Stop the program
115 *
116 DISPLAY 'Program has completed' UPON PRINTER.
117 STOP RON.
118 *
119 **********************************************************
120 ** Do the RDMS error handling - display codes, messages
121 **********************************************************
122 *
123 900-error-print.
124 *
125 *** Print errant command and error status variables
126 *
127 DISPLAY 'RDMS Command = ' room UPON PRINTER.
128 DISPLAY 'Error Status = ' err-stat UPON PRINTER.
129 DISPLAY 'Auxiliary•Info = ' aux-info UPON PRINTER.
130 *
131 *** Issue the error messages, end the thread, and exit
132 *
133 MOVE 0 TO err-end-flag.
134 PERFORM 940-geterror UNTIL no-more-msgs.
135 DISPLAY 'Error termination' UPON PRINTER.
136 IF thread-has-begun
137 MOVE ' END THREAD ;' TO room
138 ENTER MASH 'acob$rdmr' USING room, err-stat, aux-info.
139 STOP RUN.
140 *
141 *** Fill the error buffers using GETERROR
142 *
143 940-geterror.
144 MOVE ' ;' TO room.
145 ENTER MASK 'acob$rdmr' USING
146
147 PERFORM 980-print-error
148 VARYING err-s-index FROM 1 BY 1 UNTIL err-s-index > 4.
149 *
150 *** Display the error buffers
151 *
152 980-print-error.
153 IF err-s(err-s-index) NOT = SPACE
154 DISPLAY err-s(err-s-index) UPON PRINTER
155 ELSE
156 MOVE 5 to err-s-index
157 MOVE 1 TO err-end-flag.

SG 3605 C-3
D
ACOB Examples
 ACOB Examples

@PRT,S DEV.APPD/RUNACOBPROG1
FURPUR 30R2A (910822 1304:40) 1993 War 17 Wed 1210:56
ANDREWAL3644(1).APPD/RUNACOSPROG1(0)
1 @BREPT PRINT$,PRFIL.
2 @PRT,S DEV.APPD/RUNACOBPROG1
3 @ACON,SE DEV.APPD/ACOSPROG1,TPF$.ACOBPROG1
4 @MAP,S ,TPF$.ABS
5 NOT TPF$.
6 IN TPF$.ACCBPROG1
7 IN SYS$LIBPRSA.CBEP$$RSA
8 IN SYS$LISS*RSA.RDMR-ACOBDAT
9 END
10 @XQT
11 @BRITT PRINT$
12 @IPF TPF$.
13 OLD PRFIL.
14 L END,A,A,NUM
15 MOD SCR
@ACOB, SE DEV . APPD/ACOSPROG1, TPF$ .ACOBPROG1
ACOS 7R1A R 75R3J4Q 03/17/93 12:10:57 (0) 1100 ASCII COBOL 08/12/91 10:02
1100 ASCII COBOL SOURCE LISTING

1 IDENTIFICATION DIVISION.
2 PROGRAM-ID. acobprogi.
3 ENVIRONMENT DIVISION.
4 CONFIGURATION SECTION.
5 SOURCE-COMPUTER. UNISYS-2200.
6 OBJECT-COMPUTER. UNISYS-2200.
7 *
8 **********************************************************
9 *** This program uses interpretive SQL and ACOB to
10 *** illustrate simple fetches and displays from a table
11 **********************************************************
12 *
13 DATA DIVISION.
14 WORKING-STORAGE SECTION.
15 *
16 **********************************************************
17 *** Establish RDMS command error buffers and data
18 **********************************************************
19 *
20 *** Error handling variables and message buffers
21 *
22 01 room.
23 05 rlin PIC X(40) OCCURS 12 TIMES.
24 01 err-stat PIC 9(4).
25 01 aux-info PIC S1(36).
26 01 error-text.
27 05 err-s PIC X(132) OCCURS 4 TIMES.
28 01 err-s-index PIC 9(4).
29 01 thread-flag PIC 9 VALUE ZERO.
30 88 thread-has-begun VALUE 1.
31 01 cursor-flag PIC 9 VALUE ZERO.
32 88 end-of-cursor VALUE 1.
33 01 err-end-flag PIC 9 VALUE ZERO.
34 88 no-more-msgs VALUE 1.

SG 3605 D-1
ACOB Examples

35
36 *** Place for the retrieved record, set up for printing
37
38 01 record-retrieved.
39 05 rr-empno PIC 9(5).
40 05 FILLER PIC X(2) VALUE SPACES.
41 05 rr-ename PIC X(12).
42 05 FILLER PIC X(2) VALUE SPACES.
43 05 rr-job PIC X(10).
44 05 FILLER PIC X(2) VALUE SPACES.
45 05 rr-dno PIC 9(3).
46 05 FILLER PIC X(2) VALUE SPACES.
47 05 rr-hiredate PIC 9(6).
48
49 ************************************************************
50 *** BEGIN the thread, do the USE, clear the SQL buffer
51 ************************************************************
52
53 *** Start the thread
54
55 PROCEDURE DIVISION.
56 000-main.
57 MOVE 'BEGIN THREAD FOR udssrc READ TO room.
58 ENTER MASK 'acob$rdmr' USING room, err-stat, aux-info.
59 IF err-stat NOT = 0 GO TO 900-error-print.
60
61 *** Set table qualifier, clear buffer, set thread flag
62
63 MOVE 'USE DEFAULT QUALIFIER rosyschemal TO room.
64 ENTER MASK 'acob$rdmr'USING room, err-stat, aux-info.
65 IF err-stat NOT = 0 GO TO 900-error-print.
66
67 MOVE 1 TO thread-flag.
68
69 ***********************************************************
70 *** Access RDMS - DECLARE and OPEN the cursor
71 ***********************************************************
72
73 *** Declare the cursor
74
75 MOVE 'DECLARE employee CURSOR ' TO room.
76 MOVE ' SELECT empno, ename, job, hiredate, ' TO rlin(2).
77 MOVE ' dno FROM amp ;' To rlin(3).
78 ENTER MASK 'acob$rdmr' USING room, err-stat, aux-info.
79 IF err-stat NOT = 0 GO TO 900-error-print.
80
81 *** Open the cursor
82
83 MOVE 'OPEN employee TO room.
84 ENTER MASK 'acob$rdmr' USING room, err-stat, aux-info.
85 IF err-stat NOT = 0 GO TO 900-error-print.
86
87 **********************************************************
88 ** Loop thru the rows, end the thread, then do cleanup
89 **********************************************************
90
91 *** Loop to fetch a row and display it
92
93 200-loop.
94 MOVE 'FETCH employee INTO $P1, $P2, $P3, ' TO room.
95 MOVE ' $P4, $P5 ;' TO rlin(2).
96 ENTER MASK 'acob$rdmr' USING room, err-stat, aux-info,
97 rr-empno, rr-ename, rr-job, rr-hiredate, rr-dno.
98 IF err-stat = 6001 GO TO 400-drop-cursor.
99 IF err-stat NOT = 0 GO TO 900-error-print.
100

D-2 SG 3605
 ACOB Examples

101 DISPLAY record-retrieved UPON PRINTER.

102 GO TO 200-loop.
103 *
104 400-drop-cursor.
105 DISPLAY 'No more data on fetch' UPON PRINTER.
106 MOVE 'DROP CURSOR employee •' TO rcom.
107 ENTER MASM 'acob$rdmr' USING rcom, err-stat, aux-info.
108 IF err-stat NOT = 0 GO TO 900-error-print.
109 *
110 MOVE SPACES TO rcom.
111 *
112 ************************************************************
113 *** End the thread, shut down the program
114 ************************************************************
115 *
116 *** Issue the end of the thread
117 *
118 MOVE 'END THREAD •' TO rcom:
119 ENTER MASK 'acob$rdmr' USING rcom, err-stat, aux-info.
120 IF err-stat NOT = 0 GO TO 900-error-print.
121 *
122 *** Stop the program
123 *
124 DISPLAY 'Program has completed' UPON PRINTER.
125 STOP RUN.
126 *
127 **********************************************************
128 ** Do the RDMS error handling - display codes, messages
129 **********************************************************
130 *
131 900-error-print.
132 *
133 *** Print errant command and error status variables
134 *
135 DISPLAY 'RDMS Command = ' rcom UPON PRINTER.
136 DISPLAY 'Error Status = ' err-stat UPON PRINTER.
137 DISPLAY 'Auxiliary Info = ' aux-info UPON PRINTER.
138 *
139 *** Issue the error messages, end the thread, and exit
140 *
141 MOVE 0 TO err-end-flag.
142 PERFORM 940-geterror UNTIL no-more-msgs.
143 DISPLAY 'Error termination' UPON PRINTER.
144 IF thread-has-begun
145 MOVE ' END THREAD ;' TO rcom
146 ENTER MASM 'acob$rdmr' USING room, err-stat, aux-info.
147 STOP RUN.
148 *
149 *** Fill the error buffers using GETERROR
150 *
151 940-geterror.
152 MOVE 'GETERROR INTO $P1, $P2, $P3, $P4 •' TO rcom.
153 ENTER MASM 'acob$rdmr' USING rcom, err-stat, aux-info,
154 err-s(1), err-s(2), err-s(3), err-s(4).
155 PERFORM 980-print-error
156 VARYING err-s-index FROM 1 BY 1 UNTIL err-s-index > 4.
157 *
158 *** Display the error buffers
159 *
160 980-print-error.
161 IF err-s(err-s-index) NOT = SPACE
162 DISPLAY err-s(err-s-index) UPON PRINTER
163 ELSE
164 MOVE 5 to err-s-index
165 MOVE 1 TO err-end-flag.
END ACOB DIAGNOSTIC TOTALS 0 WARNING 0 MINOR 0 SERIOUS 0
FATAL 0 LEVELING
COMPILE TIME IS 3.94 SECONDS CORE: MIN= 63601 / 64000 MAX= 66560 WORDS

SG 3605 D-3
ACOB Examples

@MAP,S ,TPF$.ABS
Collector 33R1A (910808 1521:53) 1993 Mar 17 Wed 1211:02
1. NOT TPF$.
2. IN TPF$.ACOBPROG1
3. IN syssiaNsmsN.cupssNak
4. IN SYS$LIBPRSA.RDMR-ACOBDAT
5. END
AFCM status of output element is SETAFCM
Quarter-word sensitive
ADDRESS LIMITS 001000 002324 725 IBANK WORDS DECIMAL
040000 061765 9206 DRANK WORDS DECIMAL
STARTING ADDRESS 001731

SEGMENT $MAIN$ 001000 002324 040000 061765

ERU$/RLIBT12 24
MAY 90 10:37:08
m$PKT$ 040000 040005 03
AUG 90 16:10:55
MEM$ERR
C$NPADS 040006 040014 26
AUG 90 19:05:18
C$S400 040015 040027 21
DEC 90 13:37:46
C$RETO 001000 001005 26
AUG 90 18:58:21
C$DJ13 001006 001270 26
AUG 90 18:57:50
mcmgcm(commormomo 040030 040032
C$S300 $(0) 040033 040730 25
JAN 91 11:32:50
$(012) MEWERR
C$PCRLOG $(0) 040731 041030 12
AUG 91 09:47:08
CBM$$ACOB/PART-LIB-CB 12
AUG 91 10:13:59
ACOBPROG1 $(1) 001271 001743 041031 041446 17
MAR 93 12:11:01
$(3) 001744 001753 041447 041704
041705 041723
041724 042110 .
CBEP$$RSA 24
AUG 91 01:01:32
RDMR-ACOBDAT 001754 002324 $(0) 042111 061764 10
JAN 92 13:23:24
$(2) 061765 061765
Common Banks referenced
----------
C$330S 0400103 C$S116 0400105 RDMR$FTNCONT 0401074
END MAP. ERRORS: 0 TIME: 23.552 STORAGE: 016417/015211/016365/0/0130076

D-4 SG 3605
 ACOB Examples

@XQT
05010 FOSTER SALESREP 200 860714
05146 BROWN CLERK 200 871011
05234 WOODWORTH MANAGER 200 790219
05237 ROCKWELL ACCOUNTANT 100 840618
05437 MARTIN SALESREP 200 870326
05469 ADAMS SALESREP 200 811102
05630 GLASS ACCOUNTANT 100 871220
05702 TURNER CLERK 100 890515
05743 LAWSON MANAGER 100 830510
05765 JOHNSON CLERK 400 891221
05784 WILLIAMS PRESIDENT 300 781015
05896 SMITH CLERK 400 880913
05942 FORD SALESREP 200 890316
05984 HYDE MANAGER 400 850901
No more data on Patch
Program has completed
@BRKPT PRINT$

SG 3605 D-5
E
UCOB Skeleton
 UCOB Skeleton

@PRT,S AL3644*SQL.LAB3-1/RUNUCOB
FURPUR 30R2A (910822 1304:40) 1993 Mar 17 Wed 1514:48
AL3644*SQL(1).LAB3-1/RUNUCOB(0)
1 @BRKPT PRINT$,prfil.
2 @PRT,S sql.lab3-1/runucob
3 @UCOB,ES sql.lab3-1/ucobskel,tpf$.ucobskel„,no-options,narrow
4 @XQT
5 @BRKPT PRINT$
6 @IPF tpf$.
7 OLD prfil.
8 1 "END",e,e,numb
9 mod scr
@PRT,S AL3644*SQL.LAB3-1/UCOBSKEL
AL3644*SQL(1).LAB3-1/UCOBSKEL(0)
1 IDENTIFICATION DIVISION.
2 PROGRAM-ID. ucobskel.
3 ENVIRONMENT DIVISION.
4 CONFIGURATION SECTION.
5 SPECIAL-NAMES. PRINTER IS printer.
6
7 **********************************************************
8 *** This program uses embedded SQL and UCOB to
9 *** solve lab 3-1 assignment for SQL Programming
10 **********************************************************
11
12 DATA DIVISION.
13 WORKING-STORAGE SECTION.
14
15 01 thread-flag PIC 9 VALUE ZERO.
16 01 err-end-flag PIC 9 VALUE ZERO.
17 88 no-more-msgs VALUE 1.
18
19 **********************************************************
20 *** Establish RDMS command error buffers and data
21 **********************************************************
22
23 EXEC SQL BEGIN DECLARE SECTION END-EXEC.
24
25 *** Place for the retrieved record, set up for printing
26
27 01 record-retrieved.
28 05
29 05
30
31 *** Error handling variables and message buffers
32
33 01 sqlcode PIC S9(9) USAGE COMP.
34
35 01 rdmca.
36 05 error-status PIC 9(4).
37 05 aux-info PIC S9(9) USAGE BINARY.
38
39 01 err-text.
40 05 err-s PIC X(132) OCCURS 4 TIMES.
41
42 01 err-s-index PIC 9.
43 EXEC SQL END DECLARE SECTION END-EXEC.
44

SG 3605 E-1
UCOB Skeleton

45 **********************************************************
46 *** Start by setting up error handling and BEGIN thread
47 **********************************************************
48
49 PROCEDURE DIVISION.
50 000-main.
51
52 *** What to do on no-find
53
54 EXEC SQL
55
56 END-EXEC.
57
58 *** What to do if an SQL error is encountered
59
60 EXEC SQL
61
62 END-EXEC.
63
64 *** Start the thread
65
66
67 MOVE 1 TO thread-flag.
68
69 **********************************************************
70 ** Access RDMS - USE, DECLARE, OPEN and initial FETCH
71 **********************************************************
72
73 *** Set up the table qualifier
74
75 EXEC SQL
76
77 END-EXEC.
78
79 *** Declare the cursor
80
81 EXEC SQL
82
83
84
85 END-EXEC.
86
87 *** Open the cursor
88
89 EXEC SQL
90
91 END-EXEC.
92
93 *** Do an initial fetch of a row
94
95 EXEC SQL
96
97
98 END-EXEC.
99

E-2 SG 3605
 UCOB Skeleton

100 **********************************************************
101 ** Loop thru the rows, end the thread, then do cleanup
102 **********************************************************
103
104 *** Loop to display a row and get another
105
106 PERFORM UNTIL sqlcode = 100
107 DISPLAY record-retrieved
108 EXEC SQL
109 FETCH
110
111 END-EXEC
112 END-PERFORM.
113
114 **********************************************************
115 *** Handle a no-find, end thread, and terminate program
116 **********************************************************
117
118 *** Come here when a no find condition exists
119
120 400-no-find.
121
122 *** Issue the end of the thread
123
124
125 MOVE 0 TO thread-flag.
126
127 *** Stop the program
128
129 DISPLAY 'Program has completed' UPON PRINTER.
130 STOP RUN.
131

SG 3605 E-3
UCOB Skeleton

132 **********************************************************
133 ** Do the RDMS error handling - display codes & messages
134 **********************************************************
135
136 *** Come here when RDMS errors are detected
137
138 900-rdms-error.
139
140 *** Turn off error handling, print error status variables
141
142 EXEC SQL WHENEVER SQLERROR CONTINUE END-EXEC.
143
144 DISPLAY '***SQL Code = ' sqlcode UPON PRINTER.
145 DISPLAY '***Error Code = ' error-status UPON PRINTER.
146 DISPLAY '***Auxiliary Info = ' aux-info UPON PRINTER.
147
148 *** Issue the error messages, end the thread, and exit
149
150 MOVE 0 TO err-end-flag.
151 PERFORM 940-geterror UNTIL no-more-msgs.
152 DISPLAY 'Error termination' UPON PRINTER.
153 END THREAD.
154 STOP RUN.
155
156 *** Fill the error buffers using GETERROR
157
158 940-geterror.
159
160 PERFORM 980-print-error
161 VARYING err-s-index FROM 1 BY 1 UNTIL err-s-index > 4.
162
163 *** Display the error buffers
164
165 980-print-error.
166 IF err-s(err-s-index) NOT = SPACE
167 DISPLAY err-s(err-s-index) UPON PRINTER
168 ELSE
169 MOVE 5 to err-s-index
170 MOVE 1 TO err-end-flag.
@BRXPT PRINT$

E-4 SG 3605
F
UCOB Examples
 UCOB Examples

@PRT,S DEV.APPF/RUNUCOBPROG1
FURPUR 30R2A (910822 1304:40) 1993 Mar 17 Wed 1312:51
ANDREW*AL3644(1).APPF/RUNUCOBPROG1(0)
1 @BRKPT PRINT$,prfil.
2 @PRT,S dev.appf/runucobprogl
3 @UCOB,ES dev.appf/ucobprogl,tpf$.ucobprogl,,,no-options,narrow
4 @XQT
5 @BRKPT PR/NT$
6 @IPF tpf$.
7 OLD prfil.
8 1 "END",a,a,numb
9 mod scr
@UCOB,ES DEV.APPF/UCOBPROG1,TPF$.UCOBPROG1„,NO-OPTIONS,NARROW
*REMARK(CLARIFICATION) LSS-CSIM20389: Option REMARK has been previously
set
UCOB- 5R2(910816) LSS- 6R1A(910814) 930317 13:12:52
1 IDENTIFICATION DIVISION.
2 PROGRAM-ID. ucobprogl.
3 ENVIRONMENT DIVISION.
4 CONFIGURATION SECTION.
5 SPECIAL-NAMES. PRINTER IS printer.
6
7 **********************************************************
8 *** This program uses embedded SQL and UCOB to
9 *** solve lab exercise 3-1 for AL 3644.
10 **********************************************************
11
12 DATA DIVISION.
13 WORKING-STORAGE SECTION.
14
15 01 thread-flag PIC 9 VALUE ZERO.
16 01 err-end-flag PIC 9 VALUE ZERO.
17 88 no-more-msgs VALUE 1.
18
19 **********************************************************
20 *** Establish RDMS command error buffers and data
21 **********************************************************
22
23 EXEC SQL BEGIN DECLARE SECTION END-EXEC.
24
25 *** Place for the retrieved record, set up for printing
26
27 01 record-retrieved.
28 05 rr-empno PIC 9(5) VALUE ZEROS.
29 05 PIC X(2) VALUE SPACES.
30 05 rr-ename PIC X(12) VALUE SPACES.
31 05 PIC X(2) VALUE SPACES.
32 05 rr-job PIC X(10) VALUE SPACES.
33 05 PIC X(2) VALUE SPACES.
34 05 rr-dno PIC 9(3) VALUE ZEROS.
35 05 PIC X(2) VALUE SPACES.
36 05 rr-hiredate PIC 9(6) VALUE ZEROS.
37 05 PIC X(2) VALUE SPACES.
38

SG 3605 F-1
UCOB Examples

39 *** Error handling variables and message buffers

40
41 01 sqlcode PIC S9(9) USAGE COMP.
42
43 01 rdmca.
44 05 error-status PIC 9(4).
45 05 aux-info PIC S9(9) USAGE BINARY.
46
47 01 err-text.
48 05 err-s PIC X(132) OCCURS 4 TIMES.
49
50 01 err-s-index PIC 9.
51 EXEC SQL END DECLARE SECTION END-EXEC.
52
53 **********************************************************
54 *** Start by setting up error handling and BEGIN thread
55 **********************************************************
56
57 PROCEDURE DIVISION.
58 000-main.
59
60 *** What to do on no-find
61
62 EXEC SQL
63 WHENEVER NOT FOUND GO TO 400-no-find
64 END-EXEC.
65
66 *** What to do if an SQL error is encountered
67
68 EXEC SQL
69 WHENEVER SQLERROR GO TO 900-rdm3-error
70 END-EXEC.
71
72 *** Start the thread
73
74 BEGIN THREAD FOR udssrc READ.
75 MOVE 1 TO thread-flag.
76
77 **********************************************************
78 ** Access RDMS - USE, DECLARE, OPEN and initial FETCH
79 **********************************************************
80
81 *** Set up the table qualifier
82
83 EXEC SQL
84 USE DEFAULT QUALIFIER rosyschemal
85 END-EXEC.
86
87 *** Declare the cursor
88
89 EXEC SQL
90 DECLARE emp CURSOR
91 SELECT empno, ename, job, hiredate, dno
92 FROM emp
93 END-EXEC.
94
95 *** Open the cursor
96
97 EXEC SQL
98 OPEN emp
99 END-EXEC.

F-2 SG 3605
 UCOB Examples

100
101 *** Do an initial fetch of a row
102
103 EXEC SQL
104 FETCH emp INTO :rr-empno, :rr-ename,
105 :rr-job, :rr-hiredate, :rr-dno
106 END-EXEC.
107
108 **********************************************************
109 ** Loop thru the rows, end the thread, then do cleanup
110 **********************************************************
111
112 *** Loop to display a row and get another
113
114 PERFORM UNTIL sqlcode = 100
115 DISPLAY record-retrieved
116 EXEC SQL
117 FETCH emp INTO :rr-empno, :rr-ename,
118 :rr-job, :rr-hiredate, :rr-dno
119 END-EXEC
120 END-PERFORM.
121
122 **********************************************************
123 *** Handle a no-find, end thread, and terminate program
124 **********************************************************
125
126 *** Come here when a no find condition exists
127
128 400-no-find.
129
130 *** Issue the end of the thread
131
132 END THREAD.
133 MOVE 0 TO thread-flag.
134
135 *** Stop the program
136
137 DISPLAY 'Program has completed' UPON PRINTER.
138 STOP RUN.
139
140 **********************************************************
141 ** Do the RDMS error handling - display codes & messages
142 **********************************************************
143
144 *** C mo here when RDMS errors are detected
145
146 900-rdms-error.
147
148 *** Turn off error handling, print error status variables
149
150 EXEC SQL WHENEVER SQLERROR CONTINUE END-EXEC.
151
152 DISPLAY '***SQL Code = ' sqlcode UPON PRINTER.
153 DISPLAY '***Error Code = ' error-status UPON PRINTER.
154 DISPLAY '***Auxiliary Info = ' aux-info UPON PRINTER.

SG 3605 F-3
UCOB Examples

155 *
156 *** Issue the error messages, end the thread, and exit
157 *
158 MOVE 0 TO err-end-flag.
159 PERFORM 940-geterror UNTIL no-more-msgs.
160 DISPLAY 'Error termination' UPON PRINTER.
1.61 END THREAD.
162 STOP RUN.
163 *
164 *** Fill the error buffers using GETERROR
165 *
166 940-geterror.
167 GETERROR INTO :err-s(1), :err-s(2), :err-s(3), :err-s(4).
168 PERFORM 980-print-error
169 VARYING err-s-index FROM 1 BY 1 UNTIL err-s-index > 4.
170 *
171 *** Display the error buffers
172
173 980-print-error.
174 IF err-s(err-s-index) NOT = SPACE
175 DISPLAY err-s(err-s-index) UPON PRINTER
176 ELSE
177 MOVE 5 to err-s-index
178 MOVE 1 TO err-end-flag.
SIZES: CODE(EM4): 448 DATA: 1193
END UCOB- 0 ERRORS(MAJOR) 0 ERRORS(MINOR) 1 REMARK(CLARIFICATION)
@XQT
5010 FOSTER SALESREP 200 860714
5146 BROWN CLERK 200 871011
5234 WOODWORTH MANAGER 200 790219
5237 ROCKWELL ACCOUNTANT 100 840618
5437 MARTIN SALESREP 200 870326
5469 ADAMS SALESREP 200 811102
5630 GLASS ACCOUNTANT 100 871220
5702 TURNER CLERK 100 890515
5743 LAWSON MANAGER 100 830510
5765 JOHNSON CLERK 400 891221
5784 WILLIAMS PRESIDENT 300 781015
5896 SMITH CLERK 400 880913
5942 FORD SALESREP 200 890316
5984 HYDE MANAGER 400 850901
Program has completed
@BRKPT PRINT$

F-4 SG 3605