Optimization of code

Question

Hi ,
Please help me to optimize this code w.r.t performance tuning.  thanks in advance
SELECT   v~aufpl
         v~aplzl
         V~FSAVD
         V~FSEDD
         V~ISDD
         V~IEDD
         V~MGVRG
         V~GMNGA
         V~MEINH
         c~ltxa1
         c~objnr
         C~ARBID
         s~werks
         s~arbpl
         s~vornr
         s~aufnr
         s~matnr
         o~cy_seqnr
         o~dispo
         D~VERWE
         u~usr04
         u~use04
         INTO  TABLE itab
         FROM ( ( ( ( ( afvv AS v INNER JOIN afvc AS c  ON v~aufpl EQ c~aufpl and  v~aplzl eq c~aplzl )
         inner join s022 as s  on c~objnr eq s~objnr  ) INNER JOIN
         AFko  AS O ON v~AUFPL EQ O~AUFPL AND S~AUFNR EQ O~AUFNR AND S~MATNR EQ O~PLNBEZ ) INNER JOIN CRHD
         AS D ON C~ARBID EQ D~OBJID AND S~ARBPL EQ D~ARBPL AND S~WERKS EQ D~WERKS )
         inner join afvu as u on v~aufpl eq u~aufpl and v~aplzl eq u~aplzl )
                                      WHERE  s~vornr in s_vornr
                                      and    s~werks in s_werks
                                      and   s~arbpl in s_arbpl
                                      AND   D~VERWE  IN S_VERWE
                                      and    o~aufnr in s_aufnr
                                      and    o~cy_seqnr in s_seqnr
                                      and   o~dispo in s_dispo
                                      and   c~ltxa1 in s_ltxa1
                                      and  V~fsavd in p_fsavd
                                      and  V~Fsedd in p_fsedd.
Anoop

Answer 1

Please do not confuse a view which is only a definition in the data dictionary and a view which actually
contains data in the database.
My comment above was correct, if you read the statement in the comment above:
>Unless it's a materialized view (which SAP doesn't support), the performance gain through a view is more or less insignificant
>(as far as my experience goes). It may profit a little bit from datebase buffering though, but only if the view is frequently used.
two different kinds of views !!
And for the KNA1VV I see a join in the explain (DB6 database):
SELECT FROM kna1vv :
    0 SELECT STATEMENT ( Estimated Costs =  2,292E+03 [timerons] )
           1 RETURN
               2 MSJOIN
                   3 [O] FETCH KNA1
                       4 IXSCAN KNA1~0 #key columns:  1
                   5 <i> FILTER
                       6 FETCH KNVV
                           7 IXSCAN KNVV~0 #key columns:  1
same for DD03VV now on Oracle
SELECT STATEMENT ( Estimated Costs = 11 , Estimated #Rows = 201 )
        5 COUNT STOPKEY
          Filter Predicates
            4 TABLE ACCESS BY INDEX ROWID DD03L
              Estim. CPU-Costs = 4.509 Estim. IO-Costs = 0
                3 NESTED LOOPS
                  ( Estim. Costs = 10 , Estim. #Rows = 201 )
                  Estim. CPU-Costs = 133.703 Estim. IO-Costs = 10
                    1 TABLE ACCESS FULL DD02L
                      ( Estim. Costs = 2 , Estim. #Rows = 69.173 )
                      Estim. CPU-Costs = 39.023 Estim. IO-Costs = 2
                      Filter Predicates
                    2 INDEX RANGE SCAN DD03L~1
                      Search Columns: 2
                      Estim. CPU-Costs = 1.824 Estim. IO-Costs = 0
                      Access Predicates
There are materialized views, but I doubt that this is valid performance optimization in most cases.
Siegfried