how to avoid nested loops

[Former Member]

Hi ,

can any one help me , in nested loops .. how to avoid this one.. its urgent...

the code is as follows..........

OOP AT gt_indata INTO gs_indata.

• Get Period Values from Header

IF ld_first_time = gc_J.

ld_first_time = gc_N.

• Split input Header into Period Structure

• FIELD-SYMBOLS: <lfs_hex> TYPE x,

• <lfs_char> TYPE c.

• ASSIGN gc_tab TO <lfs_hex>.

ASSIGN <lfs_hex> TO <lfs_char> CASTING.

Perform split1.

CLEAR: gs_split_1.

SPLIT gs_indata AT <lfs_char> INTO

gs_split_1-head_1

gs_split_1-head_2

gs_split_1-head_3

gs_split_1-head_4

*Period 1 to 10

gs_split_1-per_1 gs_split_1-per_2 gs_split_1-per_3 gs_split_1-per_4

gs_split_1-per_5 gs_split_1-per_6 gs_split_1-per_7 gs_split_1-per_8

gs_split_1-per_9 gs_split_1-per_10

*Period 11 to 20

gs_split_1-per_11 gs_split_1-per_12 gs_split_1-per_13 gs_split_1-per_14

gs_split_1-per_15 gs_split_1-per_16 gs_split_1-per_17 gs_split_1-per_18

gs_split_1-per_19 gs_split_1-per_20

*Period 21 to 30

gs_split_1-per_21 gs_split_1-per_22 gs_split_1-per_23 gs_split_1-per_24

gs_split_1-per_25 gs_split_1-per_26 gs_split_1-per_27 gs_split_1-per_28

gs_split_1-per_29 gs_split_1-per_30

*Period 31 to 40

gs_split_1-per_31 gs_split_1-per_32 gs_split_1-per_33 gs_split_1-per_34

gs_split_1-per_35 gs_split_1-per_36 gs_split_1-per_37 gs_split_1-per_38

gs_split_1-per_39 gs_split_1-per_40

*Period 41 to 50

gs_split_1-per_41 gs_split_1-per_42 gs_split_1-per_43 gs_split_1-per_44

gs_split_1-per_45 gs_split_1-per_46 gs_split_1-per_47 gs_split_1-per_48

gs_split_1-per_49 gs_split_1-per_50

*Period 51 to 55

gs_split_1-per_51 gs_split_1-per_52 gs_split_1-per_53 gs_split_1-per_54

gs_split_1-per_55.

*Create internal table with all Periods from input file Header

CLEAR ld_tabix.

Do 56 times Varying

• WHILE w_tabix LT 56

gs_period-per FROM gs_split_1-per_1 NEXT gs_split_1-per_2.

ADD 1 TO ld_tabix.

*Break at value END in the file

IF gs_period-per = gc_END.

EXIT.

ENDIF.

APPEND gs_period TO gt_period.

• ENDWHILE.

Enddo.

*Convert Date type and and Period to real Date

LOOP AT gt_period INTO gs_period.

CASE gs_period-per(1).

WHEN gc_W.

• Week

MOVE gs_period-per+1(6) TO ld_week.

CALL FUNCTION 'WEEK_GET_FIRST_DAY'

EXPORTING

week = ld_week

IMPORTING

date = ld_date

EXCEPTIONS

week_invalid = 1

OTHERS = 2.

IF sy-subrc <> 0.

• Error Wrong Period Week

Message A000(38) with text-002.

ENDIF.

MOVE gc_2 TO gs_period-date_type.

MOVE ld_date TO gs_period-req_date.

MODIFY gt_period FROM gs_period.

WHEN gc_M.

• Month

MOVE gc_3 TO gs_period-date_type.

MOVE gs_period-per1(6) TO ld_date_alfn0(6).

MOVE 01 TO ld_date_alfn+6(2).

MOVE ld_date_alfn TO ld_date.

• Check if Date is Valid

CALL FUNCTION 'LAST_DAY_OF_MONTHS'

EXPORTING

day_in = ld_date

IMPORTING

last_day_of_month = ld_date_last

EXCEPTIONS

day_in_no_date = 1

OTHERS = 2.

IF sy-subrc <> 0.

• Error Wrong Period Month

Message A000(38) with text-003.

ENDIF.

MOVE ld_date TO gs_period-req_date.

MODIFY gt_period FROM gs_period.

WHEN OTHERS.

• Error Wrong Date Type

MESSAGE A000(38) WITH text-001.

ENDCASE.

ENDLOOP.

DESCRIBE TABLE gt_period LINES gs_max_line.

CONTINUE.

ENDIF.

*Split input Item into Field Structure

ASSIGN gc_tab TO <lfs_hex>.

ASSIGN <lfs_hex> TO <lfs_char> CASTING.

CLEAR: gs_split_2,

gt_quantity.

SPLIT gs_indata AT <lfs_char> INTO

gs_split_2-material

gs_split_2-plant

gs_split_2-version

gs_split_2-act_ind

*Period 1 to 10

gs_split_2-ant_1 gs_split_2-ant_2 gs_split_2-ant_3 gs_split_2-ant_4

gs_split_2-ant_5 gs_split_2-ant_6 gs_split_2-ant_7 gs_split_2-ant_8

gs_split_2-ant_9 gs_split_2-ant_10

*Period 11 to 20

gs_split_2-ant_11 gs_split_2-ant_12 gs_split_2-ant_13 gs_split_2-ant_14

gs_split_2-ant_15 gs_split_2-ant_16 gs_split_2-ant_17 gs_split_2-ant_18

gs_split_2-ant_19 gs_split_2-ant_20

*Period 21 to 30

gs_split_2-ant_21 gs_split_2-ant_22 gs_split_2-ant_23 gs_split_2-ant_24

gs_split_2-ant_25 gs_split_2-ant_26 gs_split_2-ant_27 gs_split_2-ant_28

gs_split_2-ant_29 gs_split_2-ant_30

*Period 31 to 40

gs_split_2-ant_31 gs_split_2-ant_32 gs_split_2-ant_33 gs_split_2-ant_34

gs_split_2-ant_35 gs_split_2-ant_36 gs_split_2-ant_37 gs_split_2-ant_38

gs_split_2-ant_39 gs_split_2-ant_40

*Period 41 to 50

gs_split_2-ant_41 gs_split_2-ant_42 gs_split_2-ant_43 gs_split_2-ant_44

gs_split_2-ant_45 gs_split_2-ant_46 gs_split_2-ant_47 gs_split_2-ant_48

gs_split_2-ant_49 gs_split_2-ant_50

*Period 51 to 55

gs_split_2-ant_51 gs_split_2-ant_52 gs_split_2-ant_53 gs_split_2-ant_54

gs_split_2-ant_55.

*Create internal table with all Periods from input file

CLEAR ld_tabix.

• WHILE w_tabix LT gs_max_line

Do gs_max_line Times Varying

gs_quantity-antal FROM gs_split_2-ant_1 NEXT gs_split_2-ant_2.

ADD 1 TO ld_tabix.

APPEND gs_quantity TO gt_quantity.

Enddo.

• ENDWHILE.

MOVE gs_split_2-material TO gs_outdata-matnr.

MOVE gs_split_2-plant TO gs_outdata-plant.

MOVE gs_split_2-version TO gs_outdata-version.

MOVE gs_split_2-act_ind TO gs_outdata-act_ind.

MOVE gt_quantity[] TO gs_outdata-quantity[].

APPEND gs_outdata TO gt_outdata.

ENDLOOP. "GT_INDATA

• Loop through Output file records and create return table for each Product

• Allocation Object.

LOOP AT gt_alloc_object INTO gs_alloc_object.

• LOOP AT gt_outdata INTO gs_outdata.

Read Table gt_outdata into gs_outdata index sy-index.

CLEAR gs_t_data.

gs_t_data-konob = gs_alloc_object-konob.

gs_t_data-matnr = gs_outdata-matnr.

gs_t_data-vrsio = gs_outdata-version.

gs_t_data-werks = gs_outdata-plant.

• Validate material and get unit of measurement from material master

SELECT SINGLE meins FROM mara

INTO gs_t_data-basme

WHERE matnr = gs_t_data-matnr.

IF sy-subrc EQ 0.

• Material exists

LOOP AT gs_outdata-quantity INTO gs_quantity.

• READ TABLE gt_outdata into gs_outdata index sy-index.

• move gs_outdata-quantity to gs_quantity-antal.

MOVE gs_quantity-antal TO gs_t_data-kcqty.

• Get period

SORT gt_period.

READ TABLE gt_period INTO gs_period INDEX sy-tabix .

IF SY-SUBRC = 0.

gs_t_data-spwoc = gs_period-per+1(6).

ENDIF.

• Check restrictions

IF gs_t_data-werks IN gt_werks AND

gs_t_data-matnr IN gt_matnr AND

gs_t_data-spwoc IN gt_spwoc.

APPEND gs_t_data TO t_data.

ENDIF.

ENDLOOP. "gs_OUTDATA-QUANTITY

ENDIF.

• ENDLOOP. "GT_OUTDATA

ENDLOOP. "GT_ALLOC_OBJECT

ENDIF.

ENDIF.

WHEN OTHERS.

• No more info structures implemented

ENDCASE.

form split1.

endform.

[Former Member]

hi

You can use read statement using binary search instead of using nested loop in some places at ur program. but don’t forget to sort the table which u r reading before reading that.

One more thing in side the loop u used select single *. But u can use select with required fields out side of the loop in to itab, then read that itab in side the loop.

Reward if it dose.

Thanks

Siva Kumar

[Former Member]

You can not avoid nested loops, because it is often necessary to compare to internal tables,

1. One table must be the outher table and must be fully processed,

Each each line of the outer table:

You must not process a full inner table.

But only one line of the inner table, then use READ ... BINARY SEARCH' and SORT this table once outside of the outer table!

Or a few lines of the inner table, then use READ BINARY SEARCH and LOOP ... FROM INDEX with EXIT condition, which is very important.

Read this blog, especially the last section:

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

Siegfried

[Former Member]

For better performance go for Parellel cursor technique or fetch data in internal tables using Select .... FOR all enteries....

Regards

Naveen Gupta

[Former Member]

see this link:

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

how to avoid nested loops in this program to improve the performence

Reward points if useful.

[Former Member]

Hi,

using parllel cursor technique you can avoid nested loop.

for example:-

data: l_index like sy-index.

loop at itab.

read table itab2 with key ...........

if sy-subrc = 0.

l_index = sy-index.

do.

"Here write the code what you want".

again go for read.

l_index = l_index + 1

read table itab2 index l_index.

if compare the all fields which you are using while

reading first time from the table itab2.

else.

exit.

endif.

enddo.

clear l_index.

endif.

endloop.

Reward if useful.

Thanks,

Sreeram.

[Former Member]

Hi,

The exact Parllel cursor technique is :

SORT : it_mara BY matnr ,

it_makt BY matnr .

i = 1.

LOOP AT it_mara INTO wa_mara.

LOOP AT it_makt INTO wa_makt FROM i.

IF wa_makt-matnr <> wa_mara-matnr.

i = sy-tabix.

EXIT.

ENDIF.

write 😕 wa_mara-matnr , wa_makt-MAKTX .

MOVE wa_mara-matnr TO itab-matnr.

MOVE wa_makt-maktx TO itab-maktx.

append itab .

ENDLOOP.

ENDLOOP.

PS: REward points if helpfull

Regards

Naveen Gupta