KONV - alternative selection

[vijay_kumar134]

Hi  Friends,

Please guide me for alternative for following code. Becuase KONV is a cluster table- fetching date for more time.

select knumv
           kposn
           stunr
           zaehk
           kwert
           kschl from konv
                 into table git_konv1
                 for all entries in git_vbrk
                 where knumv = git_vbrk-knumv
                   and kposn = git_vbrk-posnr
                   and kschl = c_jexp.

select knumv
           kposn
           stunr
           zaehk
           kwert
           kschl from konv
                 into table git_konv
                 for all entries in git_vbrk
                 where knumv = git_vbrk-knumv
                   and kposn = git_vbrk-posnr
                   and kschl in (c_znsr , c_zr00 , c_zr01 , c_za03 , c_za04 , c_za05 , c_za06 , c_za10 , c_zscg , c_zstc , c_zrtc ,
                                 c_zptc , c_kf00 , c_zpac , c_z0cd , c_jass , c_zocd ,c_z007 , c_zoth , c_zwhc , c_zwhs ,c_zk01 , c_zk02 ).
                                
Thanks.

Viji.

[Sijin_Chandran]

HI Vijay ,

Why you have  wrote SELECT statements twice for where clause and kschl = c_jexp ?

Instead write a single for both and Process  kschl = c_jexp separately while processing the internal table data in loop . ( Using READ STATEMENT ).

And FOR ALL ENTRIES is the only way for joining a CLUSTER table.

Always try to avoid unnecessary select statements .

[Former Member]

You are using the two left most key fields of KONV in your selection - KNUMV and KPOSN and also that you are restricting by condition types KSCHL, so the select is already optimized to a good extent. If it is taking time it should be because your selection must also be large in git_vbrk internal table.

If you are only selecting pricing records (and NOT tax conditions) you can also add additional WHERE clause KAPPL = 'V'

Also why are you making two SELECTs? You can merge the first select into the second and later you can filter the JEXP records separately. That is more efficient