parallel cursor.

[Former Member]

hi,

Please explain with exactly how the parallel cursor means (Whether only using INDEX).

I have a scenario like this

three internal tables i am using loop at itab WHERE,

how to make the same using parallel cursor, which will improve the performance.

chk this i want similar scenario.

tables:ekko,ekpo.

*parameter: p_ebeln like ekko-ebeln.

select-options: p_ebeln for ekko-ebeln.

data: i type i.

data: itab1 type table of ekko with header line,

itab2 type table of ekpo with header line,

itab3 type table of ekbe with header line.

select * from ekko into table itab1 where ebeln in p_ebeln.

select * from ekpo into table itab2 where ebeln in p_ebeln.

select * from ekbe into table itab3 where ebeln in p_ebeln.

*i = 1.

loop at itab1.

write / 'loop 1'.

loop at itab2 where ebeln = itab1-ebeln.

write / 'loop2'.

loop at itab3 ebeln = itab2-ebeln ebelp = itab2-ebelp.

write / 'loop3'.

endloop.

endloop.

endloop.

regards,

Prabhu

Points assured.

[Former Member]

Try the below code. Hope it helps.

sort iekpo by ebeln ebelp.
sort iekbe by ebeln ebelp.

loop at iekko.
 clear: lv_inx.
 read table iekpo with key ebeln = iekko-ebeln binary search.
 if sy-subrc = 0.
   lv_indx = sy-tabix.
   loop at iekpo from lv_indx.
     if iekko-ebeln = iekpo-ebeln.
       clear: lv_indx1.
       read table iekbe with key ebeln = iekpo-ebeln
                                            ebelp = iekpo-ebelp
                                            binary search.
       if sy-subrc = 0.
          lv_indx1 = sy-tabix.
          loop at iekbe from lv_indx1.
             if iekpo-ebeln = iekbe-ebeln and
                iekpo-ebelp = iekbe-ebelp.
                 <<< do ur logic>>
             else.
                 clear: lv_indx1.
                 exit.
              endif.
            endloop.
     else.
       clear: lv_indx.
       exit.
     endif.
   endloop.
 endif.
endloop.

Thanks,

Balaji

[Former Member]

Hi Balaji,

The code is fine if have one entry in itab2 for itab1 and one entry for itab3 in itab1 then you code is fine.

But if we have 2 record(itab2) for 1 entry in itab1 and so on it doesn't work as per the requirement

regards,

Prabhu

[Former Member]

Try the below code.

sort iekpo by ebeln ebelp.
sort iekbe by ebeln ebelp.
 
loop at iekko.
 write:/ iekko.
 clear: lv_inx.
 read table iekpo with key ebeln = iekko-ebeln binary search.
 if sy-subrc = 0.
   lv_indx = sy-tabix.
   loop at iekpo from lv_indx.
     if iekko-ebeln = iekpo-ebeln.
       write:/ iekpo.
       clear: lv_indx1.
       read table iekbe with key ebeln = iekpo-ebeln
                                            ebelp = iekpo-ebelp
                                            binary search.
       if sy-subrc = 0.
          lv_indx1 = sy-tabix.
          loop at iekbe from lv_indx1.
             if iekpo-ebeln = iekbe-ebeln and
                iekpo-ebelp = iekbe-ebelp.
                write:/ iekbe.
                 <<< do ur logic>>
             else.
                 clear: lv_indx1.
                 exit.
              endif.
          endloop.
      endif.
   endloop.
 endif.
endloop.

Hope this helps.

Thanks

Balaji

[Former Member]

hi balaji,

Thanx, whether this will be more efficient then using Loop at with where condition.

regards,

Prabhu

[Former Member]

This will be more efficient when the number of records in itab1, itab2 and itab3 are more.

This will reduce lot of processing time.

Rwd points if helpful.

Thanks

Balaji

[Former Member]

[www.saptechnical.com/Tutorials/ABAP/ParallelCursor.htm]

Got to this link what u require is there perfect with examples

and about the performances while using parallel cursor

regards,

priya.

[rodrigo_paisante3]

Hi,

see this link.

/people/rob.burbank/blog/2006/02/07/performance-of-nested-loops

Rob shows the performance comparing 4 routines, including nested loop, indexed and parallel.

and this.

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

Siegfried explain how parallel works. Good job

Regards!