The following sample code explains the usage of index looping and shows the difference between nested loops and index looping:
REPORT ZDEMOINDEXLOOPING.
*Internal tables and structures:
DATA:
lt_bkpf TYPE STANDARD TABLE OF bkpf,
lt_bseg TYPE STANDARD TABLE OF bseg,
lst_bkpf TYPE bkpf,
lst_bseg TYPE bseg.
*Variables:
DATA:
gv_stime TYPE i, "Start Time
gv_etime TYPE i, "End Time
gv_tdiff TYPE i, "Time Difference
gv_counter TYPE i, "Record Counter
gv_tabix TYPE sy-tabix. "Tabix Position
*------------------------------------*
*DUMMY RETRIEVAL: *
*------------------------------------*
*Dummy retrieval for demo purposes.
*Data at header level.
SELECT *
FROM bkpf
INTO TABLE lt_bkpf
UP TO 10 ROWS.
IF sy-subrc EQ 0.
* Sort and delete adjacent duplicates not necessary as
* we are sorting and deleting by primary keys. However, we
* will leave like that to remember that its a best practice
* to sort and delete adjacents duplicates when using for
* all entries.
SORT lt_bkpf BY bukrs belnr gjahr.
DELETE ADJACENT DUPLICATES FROM lt_bkpf COMPARING bukrs belnr gjahr.
ENDIF.
*If the table is not empty, proceed. Please note that
*this is necessary when doing for all entries because
*if the table 'LT_BKPF' is empty, the following retrieval
*will retrieve all the entries from table BSEG.
IF lt_bkpf[] IS NOT INITIAL.
* Data at item level.
SELECT *
FROM bseg
INTO TABLE lt_bseg
FOR ALL ENTRIES IN lt_bkpf
WHERE bukrs EQ lt_bkpf-bukrs
AND belnr EQ lt_bkpf-belnr
AND gjahr EQ lt_bkpf-gjahr.
IF sy-subrc EQ 0.
* Sorting not required but its just for
* information purposes.
SORT lt_bseg BY bukrs belnr gjahr.
ENDIF.
ENDIF.
*------------------------------------*
*NESTED LOOPS: *
*------------------------------------*
*Get the starting time.
GET RUN TIME FIELD gv_stime.
*Nested Loops:
*Loop at header level.
LOOP AT lt_bkpf INTO lst_bkpf.
* Loop at item level.
LOOP AT lt_bseg INTO lst_bseg
WHERE bukrs EQ lst_bkpf-bukrs
AND belnr EQ lst_bkpf-belnr
AND gjahr EQ lst_bkpf-gjahr.
* Do some processing.
gv_counter = gv_counter + 1.
ENDLOOP.
ENDLOOP.
*Get the ending time.
GET RUN TIME FIELD gv_etime.
*Get the time difference, which will be the execution time
*the code was executed.
gv_tdiff = gv_etime - gv_stime.
*Output the time difference.
WRITE:/1(61) sy-uline.
WRITE: /1 sy-vline, 2(40) 'Taking Taken Using Nested Loops:', gv_tdiff,
41 sy-vline,
50(10) gv_tdiff,
61 sy-vline.
*------------------------------------*
*INDEX LOOPING: *
*------------------------------------*
CLEAR:
gv_counter.
*Get the starting time.
GET RUN TIME FIELD gv_stime.
*Nested Loops:
*Loop at header level.
LOOP AT lt_bkpf INTO lst_bkpf.
* We will check if we have the header record
* at item level.
READ TABLE lt_bseg
TRANSPORTING NO FIELDS
WITH KEY bukrs = lst_bkpf-bukrs
belnr = lst_bkpf-belnr
gjahr = lst_bkpf-gjahr
BINARY SEARCH.
* If we have a match, get the tabix position
* at which we found the header level record in
* the item level table.
IF sy-subrc EQ 0.
* Get the tabix position of the line found
* in 'LT_BSEG'. The tabix position will be
* used to start the next loop at that position.
gv_tabix = sy-tabix.
* Loop at item level.
* Here, we will start looping in the table 'LT_BSEG'
* using the tabix position found above.
LOOP AT lt_bseg INTO lst_bseg
FROM gv_tabix.
* Here, we will check if the current item level
* record still match the header level record. If not,
* means that we can exit the loop and take the next
* header level record.
IF lst_bkpf-bukrs NE lst_bseg-bukrs
OR lst_bkpf-belnr NE lst_bseg-belnr
OR lst_bkpf-gjahr NE lst_bseg-gjahr.
* Clearing necessary items.
CLEAR:
gv_tabix,
lst_bseg.
* Okay, go out of the loop.
EXIT.
ENDIF.
* If the IF statement above has failed, means that
* that line at header level matches the line at item
* level, do some processing.
* Do some processing.
gv_counter = gv_counter + 1.
* Clearing necessary items.
CLEAR:
lst_bseg.
ENDLOOP.
ENDIF.
* Clearing necessary items.
CLEAR:
lst_bkpf.
ENDLOOP.
*Get the ending time.
GET RUN TIME FIELD gv_etime.
*Get the time difference, which will be the execution time
*the code was executed.
gv_tdiff = gv_etime - gv_stime.
*Output the time difference.
WRITE:/1(61) sy-uline.
WRITE: /1 sy-vline, 2(40) 'Taking Taken Using Index Looping:', gv_tdiff,
41 sy-vline,
50(10) gv_tdiff,
61 sy-vline.
WRITE:/1(61) sy-uline.
Serkan AKKAVAK
Computer Engineer
ABAP Developer & SAP S/4 HANA Logistics Consultant
Contact : serkurumsal@yandex.com
Social Plugin