loop with in loop

[Former Member]

hi experts,

can any body tell me how can a loop with in loop effect the performance of the abap program.

can we use like that .

[RaymondGiuseppi]

Usually "loop in loop" looks like

LOOP AT header_itab.
  LOOP AT item_itab WHERE num = header_itab-num.
  ENDLOOP.
ENDLOOP.

With standard (eg index) tables the item_tab can be fully read to select records at each pass. Only way to optimize is to have declared this table as a SORTED TABLE with the header number as a key. In this case the internal loop is optimized. - This is written in online abap documentation.

Regards

[Former Member]

thanks a lot for ur reply.

[Former Member]

loop at itab1. ( n records)

loop at itab2. (m records)

endloop.

endloop.

for each record of itab1. entire itab2 is looping so finally inner loop is going to run n * m times.

so it is very poor performance

to avoid this u need to use read statement

[Former Member]

thanks a lot for ur reply.

[Former Member]

If the question is already answered then you forgot to give points!

There is general assumption that you have either

loop at itab1. ( n records)

loop at itab2. ( a records, a constant)

or

loop at itab1

loop at itab2 where key = (m records in itab2, where m grows if n grows)

In the first case you have no performance problem, it scales with n.

In the second case, you have a performance problem if itab2 is standard table. But you can avoid it, if you use sorted tables.

You can also avoid it for standard tables by a workaround

read binary search, loop from index, exit

For details on the workaround see

Measurements on internal tables: Reads and Loops:

/people/siegfried.boes/blog/2007/09/12/runtimes-of-reads-and-loops-on-internal-tables

For background information on nonlinearity see

/people/siegfried.boes/blog/2007/02/12/performance-problems-caused-by-nonlinear-coding

Siegfried