ABAP BEST PERFORMANCE
AUTHOR: SUSHANT SINGH CHHATRYA (11066)
SAP BRF+ CONSULTANT
Purpose:
- To familiarize ABAP developers with best performance guidelines for writing ABAP code.
Scope:
- Include several points highlighting the pros and cons of different syntax and coding
methods.
- Assist ABAP developers in enhancing code efficiency.
Benefits:
- Learn Best practices of ABAP.
- Understand the fundamentals of space and time complexity.
JULY 22, 2024
INVENIOLSI
SUSHANT SINGH 0
Introduction:
In computer science, there are many ways to solve a problem using
different methods.
It's important to compare these methods to see which one works best.
- Time complexity: tells us how fast an algorithm runs as the input
size grows.
- We want algorithms that run quickly, so our apps don't slow down.
- E.g. As ABAPer, how we can enhance our loops to minimise
execution time.
- Space complexity: tells us how much memory an algorithm need.
- We aim to use memory efficiently to avoid running out of space on
our systems.
- E.g. How we can remove all the wanted data declaration and use best
practice instead of hard coding.
ABAP: Best performance Guidelines:
SUSHANT SINGH 1
- CORRESPONDING USE BELOW:
- Below is structure declaration
- Now below is the select query will changed sequence
- Now, with this syntax it will not throw error, but lowers the performance.
- SAP will go to the field-by-field mapping with Corresponding.
- Without its SAP will go by default sequence wise (Best Performance).
Traditional DB: Oracle, DB2, Sybase. (Cannot handle large data)
In Join – Load to multiple tables in a database
SUSHANT SINGH 2
For all entries – Load to only one table (We write individual queries)
- Below here at a time load to only one db. table:
- Goto -> System -> Status -> Check DB
SUSHANT SINGH 3
- In case of HANA DB (HDB) we can easily work with joins.
Note: It will fetch all the data from for all entries dependent table,
which is wrong:
- Below lt_data is empty, and we have not check it.
- Leads to fetch full data.
- Wrong o/p.
----------------------------------------------------------------------------------------------
- Always remember to SORT the table first cause binary search
work on sorted tables.
- Please refer to Searching algorithms for more information.
SUSHANT SINGH 4
- Searching and Sorting concepts comes under DSA (Data
structure and algorithms)
- Header table: LT_DATA
- Item table: LT_DATA1.
1. Nested Loops: Looping 8*8*8 times = 64 times
SUSHANT SINGH 5
2. Enhance it with parallel cursor: only 8 times
Note Below:
Sy-tabix: Return current iteration of the current loop.
I – Integer – Complete data type – No need to specify the length.
- Work for Index: 1, 4, 6
- Whenever order no. is not matching, we exit from the loop, in that way we save our
iterations.
SUSHANT SINGH 6
- Always fetch the data one time and then apply loop or read table
to the output.
If: Checks all the conditions one by one, Less performance.
Case: It will directly jump to the condition.
SUSHANT SINGH 7
Index in SAP: Used to directly points to the record
Primary Index: Created automatically by Primary key in the table.
Secondary Index: We can create
- Below is screenshot to check and create index:
SUSHANT SINGH 8
- Below, Order no. is the Primary key, and we are fetching based on
it: It is good practice:
Here, below, order date is not Primary key, and it is our requirement, it
will affect performance.
Note: On creating Secondary index, it will store as a separate
repository in database table:
So, it occupies space on the DB!
SUSHANT SINGH 9
➔ Always Check for Secondary Index first, if provided by SAP.
➔ Check whether there is need for secondary index as it increases
space complexity.
Happy Learning
SUSHANT SINGH 10
