How to make this code go faster?

Question

Hi,
I have this code and its really slow. My question is how to make it faster since its taking about 10 hours.
Code:
  loop at ti_equi.
    select matnr sernr lief_nr kunde datum uzeit bwart posnr
appending corresponding fields of table ti_ser01
      from z_objk_ser01
     where sernr eq ti_equi-sernr
       and matnr eq ti_equi-matnr.
  endloop.
  delete ti_ser01 where not ( kunde in s_kunnr and
                              bwart = '631' ).
  sort ti_ser01 by matnr sernr datum descending uzeit descending.
  delete adjacent duplicates from ti_ser01 comparing matnr sernr.
What this code does is fetch all these fields and then just keeps the one that's been modified at last.
I have created an index by sernr and matnr in table objk which makes it just a little bit faster.
Any helpful answer will be appreciated.
Regards,
Roberto
Edited by: Julius Bussche on Jan 27, 2009 4:29 PM
Code tags added. Please use more meaningfull subject titles

Answer 1

Hi roberto
please avoid select statment inside loop.
use code
data: index type sy-tabix.
data : ti-ser011 type standard table of zobjk-ser01 with header line.
select matnr sernr lief_nr kunde datum uzeit bwart posnr
      from z_objk_ser01 into table ti_ser01.
sort ti_ser01 by sernr matnr.
sort ti_equi by sernr matnr.
index = 1.
loop at ti_equi.
    loop at ti_ser01 from index
     if  ti_ser01-sernr = ti_equi-sernr
          and ti_ser01-matnr = ti_equi-matnr.
        move-corresponding ti_ser01 to ti_ser011.
        append ti_ser011.
        index = sy-tabix.
        exit.  
     endif.
    endloop.
endloop.
here i have used one more internal table ti_ser011.
you can delete specific record from ti_ser01 internal table on condition. at that time u don't need to create any other internal table like ti-ser011.
in place of loop select statement---endloop.
you can use this code. I think it will help you.
please find if it is udeful.