02.

INDEXING
ADVANCED DATABASE MANAGEMENT SYSTEMS
ICT 331-2
INTRODUCTION

 Indexes used to speed up record retrieval in response to

certain search conditions.
 Index structures provide secondary access paths.
 Any field can be used to create an index.
 Multiple indexes can be constructed
 Most indexes based on ordered files
 Tree data structures organize the index
TYPES OF SINGLE-LEVEL ORDERED INDEXES
 Ordered index similar to index in a textbook.
 Indexing field (attribute)
 Index stores each value of the index field with list of
pointers to all disk blocks that contain records with that
field value
 Values in index are ordered
 Primary index
 Specified on the ordering key field of ordered file of
records
TYPES OF SINGLE-LEVEL ORDERED INDEXES (CONT’D.)

 Clustering index
 Used if numerous records can have the same value for
the ordering field
 Secondary index
 Can be specified on any nonordering field
 Data file can have several secondary indexes
PRIMARY INDEXES
 Ordered file with two fields
 Primary key, K(i)
 Pointer to a disk block, P(i)
 One index entry in the index file for each block in the data
file
 Indexes may be dense or sparse
 Dense index has an index entry for every search key
value in the data file
 Sparse index has entries for only some search values
PRIMARY INDEXES (CONT’D.)

Primary index on the ordering key field of the file

PRIMARY INDEXES (CONT’D.)

 Major problem: insertion and deletion of records

 Move records around and change index values
 Solutions
 Use unordered overflow file
 Use linked list of overflow records
SOLUTIONS: USE UNORDERED OVERFLOW FILE

 Create an overflow file to store new records that cannot fit

into the main ordered file without disrupting its order.
 When an insertion occurs, the new record is placed in the
overflow file instead of adjusting the main file.
SOLUTION: LINKED LIST OF OVERFLOW

 Use a linked list structure to link overflow records to their

original block in the main file.
 Each block in the main file has a pointer to its
corresponding overflow records.
CLUSTERING INDEXES

 Clustering field
 File records are physically ordered on a nonkey field
without a distinct value for each record
 Structure of the Ordered File
 Same type as clustering field
 Disk block pointer
CLUSTERING INDEXES

 The clustering index has an entry for each distinct value of

the clustering field.
 There can be only one clustered index per table.
 Blocks of fixed size are reserved for each value of the
clustering field to avoid physical reordering during
insertion and deletion.
 A clustering index on the
Dept_number ordering
nonkey field of an
EMPLOYEE file

To locate a record:
• Search for the clustering field
value (K(i)) in the Index File.
• Use the Block Pointer (P(i)) to
access the block in the Data
File.
• Search for the record within
the block.
SECONDARY INDEXES
 provides a secondary means of accessing a file for which
some primary access already exists.
 Ordered file with two fields
 Indexing field, K(i)
 Block pointer or record pointer, P(i)
 Usually need more storage space and longer search time
than primary index
 Improved search time for arbitrary record
 Dense secondary index
(with block pointers) on a
SECONDARY INDEXES (CONT’D.)
nonordering key field of a
file.

To retrieve a record:
• Find the Index Field Value in
the Index File.
• Use the Block Pointer to
access the corresponding
block in the Data File.
• Search for the desired record
within the block using the
secondary key field.
TYPES OF SINGLE-LEVEL ORDERED INDEXES (CONT’D.)

Table 1 Types of indexes based on the properties of the indexing field

Table 2 Properties of index types

MULTILEVEL INDEXES

 Designed to greatly reduce remaining search space as

search is conducted
 Reduces the search space by the blocking factor (𝑏𝑓𝑟),
also called the fan-out ( ).
 represents the number of entries in a single block and is
larger than 2.
 Searching a multilevel index requires approximately block
accesses.
 Faster than binary search when 𝑓𝑜>2.
MULTILEVEL INDEXES

 Because a single-level index is an ordered file, we can

create a primary index to the index itself ; in this case, the
original index file is called the first-level index and the
index to the index is called the second-level index.
 We can repeat the process, creating a third, fourth, ..., top
level until all entries of the top level fit in one disk block
 A multi-level index can be created for any type of first-
level index (primary, secondary, clustering) as long as the
first-level index consists of more than one disk block
MULTILEVEL INDEXES

 Index file
 Considered first (or base level) of a multilevel index
 Second level
 Primary index to the first level
 Third level
 Primary index to the second level
 A two-level primary index
resembling ISAM (indexed
sequential access method)
organization

[Link] the two-level index

structure in the diagram,
how would you locate the
record with a primary key
of 46?
 If a new record with a primary key of 90 needs to be inserted, how would the two-level index structure be updated?

A two-level primary index

resembling ISAM (indexed
sequential access method)
organization

[Link] the two-level index

structure in the diagram,
how would you locate the
record with a primary key
of 46?
DYNAMIC MULTILEVEL INDEXES USING B-TREES AND B+ -
TREES
 Tree data structure terminology
 Tree is formed of nodes
 Each node (except root) has one parent and zero or
more child nodes
 Leaf node has no child nodes
 Unbalanced if leaf nodes occur at different levels
 Nonleaf node called internal node
 Subtree of node consists of node and all descendant
nodes
 DYNAMIC MULTILEVEL INDEXES USING B-TREES AND B+ -
TREES
 Because of the insertion and deletion problem, most multi-
level indexes use B-tree or B+-tree data structures, which
leave space in each tree node (disk block) to allow for new
index entries
 These data structures are variations of search trees that
allow efficient insertion and deletion of new search values.
 In B-Tree and B+-Tree data structures, each node
corresponds to a disk block
 Each node is kept between half-full and completely full
TREE DATA STRUCTURE

A tree data structure that shows an unbalanced tree

SEARCH TREES AND B-TREES

 Search tree used to

guide search for a
record
 Given value of one
of record’s fields

A node in a search tree with pointers to subtrees below it

SEARCH TREES AND B-TREES (CONT’D.)

 Algorithms
necessary for
inserting and
deleting search
values into and
from the tree
A search tree of order p = 3
B-TREES

 Provide multi-level access structure

 Tree is always balanced
 Space wasted by deletion never becomes excessive
 Each node is at least half-full
 Each node in a B-tree of order p can have at most p-1
search values
B-TREE
B+ -TREES
 Data pointers stored only at the leaf nodes
 Leaf nodes have an entry for every value of the search
field, and a data pointer to the record if search field is a
key field
 For a nonkey search field, the pointer points to a block
containing pointers to the data file records
 Internal nodes
 Some search field values from the leaf nodes repeated
to guide search
B+ -TREES (CONT’D.)

The nodes of a B+-tree (a) Internal node of a B+-tree with q−1 search values (b)
Leaf node of a B+-tree with q−1 search values and q−1 data pointers
SEARCHING FOR A RECORD WITH SEARCH KEY FIELD VALUE
K, USING A B+ -TREE

Algorithm : Searching for a

record with search key field
value K, using a B+ -Tree
INDEXES ON MULTIPLE KEYS

 Multiple attributes involved in many retrieval and update

requests
 Composite keys
 Access structure using key value that combines
attributes
 Partitioned hashing
 Suitable for equality comparisons
INDEXES ON MULTIPLE KEYS (CONT’D.)

 Grid files
 Array with one
dimension for
each search
attribute

Example of a grid array on Dno and Age attributes

OTHER TYPES OF INDEXES
 Hash indexes
 Secondary structure for file access
 Uses hashing on a search key other than the one used
for the primary data file organization
 Index entries of form (K, Pr) or (K, P)
 Pr: pointer to the record containing the key
 P: pointer to the block containing the record for that
key
BITMAP INDEXES

 Used with a large number of rows

 Creates an index for one or more columns
 Each value or value range in the column is indexed
 Built on one particular value of a particular field
 Array of bits
 Existence bitmap
 Bitmaps for B+ -tree leaf nodes
FUNCTION-BASED INDEXING
 Value resulting from applying some function on a field (or
fields) becomes the index key
 Introduced in Oracle relational DBMS
 Example
 Function UPPER(Lname) returns uppercase
representation

 Query
SOME GENERAL ISSUES CONCERNING INDEXING

 Physical index
 Pointer specifies physical record address
 Disadvantage: pointer must be changed if record is moved
 Logical index
 Used when physical record addresses expected to change
frequently
 Entries of the form (K, Kp)
ADDITIONAL ISSUES RELATED TO STORAGE OF RELATIONS
AND INDEXES
 Enforcing a key constraint on an attribute
 Reject insertion if new record has same key attribute as
existing record
 Duplicates occur if index is created on a nonkey field
 Fully inverted file
 Has secondary index on every field
 Indexing hints in queries
 Suggestions used to expedite query execution
ADDITIONAL ISSUES RELATED TO STORAGE OF RELATIONS
AND INDEXES (CONT’D.)

 Column-based storage of relations

 Alternative to traditional way of storing relations by row
 Offers advantages for read-only queries
 Offers additional freedom in index creation
THANK YOU!