Write a comment

Performance for LOOP with secondary key explained

Hello Björn and thx for this article. I've been working on a similar documentation for my department and also stumbled across the weirdly slow access using the secondary key (in Loops/Filter or VALUE expressions).

After some considerable digging, I found the solution in the ABAP-Keyword Documentation: https://help.sap.com/doc/abapdocu_752_index_htm/7.52/en-US/abenitab_key_secondary_update.htm

The secondary key of a table does not automatically initialized until it is used for the first time. So your time for the Loop with key includes this initilization for the secondary key.

If you force an initialization manually beforehand with the method cl_abap_itab_utilities=>flush_itab_keys( CHANGING itab = mt_data ). The loop itself will be much faster:

Without initialization: Loop with key - 2765 : 0.1345390 Filter and Lines - 2765 : 0.0005960

With: Loop with key - 2685 : 0.0003340 Filter and Lines - 2685 : 0.0004980

This behaviour is only really relevant in test programs like this, in real life code it's more about the hundreds or thousands of loops after the first one, so the initialization does not matter that much.

Best regards, Ralf