How can i improve this code ?

Question

DATA: t_bkpf TYPE bkpf.
DATA: t_bseg type bseg_t.
DATA: wa_bseg like line of t_bseg.
select * from bkpf into t_bkpf where document type ='KZ' and bldat in s_bldat.
select single * from bseg into wa_bseg where belnr = t_bkpf-belnr.
append wa_bseg to t_bseg.
endselect.
loop at t_bseg into wa_bseg.
  at new belnr.
     if wa_bseg-koart EQ 'K'.
        // pick vendor wrbtr
     else
       // pick other line item's wrbtr
     endif.
  endat.
endloop.
i am guessing my select statements arnt efficient performance wise, secondly in the loop when i use  'at new belnr' it aint showing my any values  whereas i get all the vendors(KOART EQ 'K') when i dont use 'at new belnr' .
why is this so and how can i make the code efficient ?
Thanks..
Shehryar

Answer 1

Hi,
1.Dont read all the fields from the table unless it is required in your program.
This will tremendously improve your performance.
2.Make use of the key fields of the table wherever possible
In your scenario you could use the fields BUKRS,BELNR,GJAHR of the table BKPF in the WHERE Clause rather than
other fields.This will improve your performance a lot..
3.As BSEG is a cluster table it will cause performance problem in most cases.So try to read
the required fields from it rather than reading all the fields.Again Make use of the key fields in the WHERE Clause
here too to improve the performance..
4.Remove SELECT..ENDSELECT and replace it with FOR ALL ENTRIES to improve the performance.
Cheers,
Abdul Hakim
Mark all useful answers..