Problem in moving data of an internal table to other internal tables

[Former Member]

My internal table ITAB contains two columns NODE and VALUE.

The contents are as follows

NODE VALUE

A 1

B 2

C 3

D 4

E 5

F 6

D 7

E 8

F 9

G 10

H 11

I 12

G 13

H 14

I 15

G 16

H 17

I 18

J 19

K 20

L 21

J 22

K 23

L 24

My requirement is I need to move the values to JTAB such that JTAB should have columns as D E F and should be having the following values filled in the columns.

JTAB

D E F

4 5 6

7 8 9

Similarly I need to move the values to KTAB such that KTAB should have columns as G H I and should be having the following values filled in the columns.

KTAB

G H I

10 11 12

13 14 15

16 17 18

Similarly I need to move the values to NTAB such that NTAB should have columns as G H I and should be having the following values filled in the columns.

NTAB

J K L

19 20 21

22 23 24

Please help. Any help is higly appreciated. if dont want to put several if conditions as it will effect performance. is there a simple way to do this

[Former Member]

In looking at your example, the requirement is not clear to me. But your statement about multiple "IF" statements affecting performance is completely wrong.

Unless you're running on a 8088 chip from 1983, the CPU time to process multiple IF statements is not measurable by any instruments within 1000 miles of where I am located.

[Former Member]

U can go ahead with using if or case with even worrying about performance, if u still have doubts write ur code using if statements and execute ur program using se30 and analyse its performance there.

[former_member194669]

Check for logic creation of JTAB

    i_ref_descr1 ?= cl_abap_tabledescr=>describe_by_data( jtab ).
    i_details1[] = i_ref_descr1->components[].
    sort itab by node.
    loop at itab.
        read table i_details1 into wa_details1
                              with key name = itab-node.
        if sy-subrc eq 0.
          assign component sy-tabix of structure jtab to <fs>.
          move itab-value to <fs>.
          append jtab.
        endif.
    endloop.

After this code your jtab looks like

D E F
4
7
  5
  8
    6
    9

But still you need to Jtab look like this way

D E F
4 5 6
7 8 9

I am seeing that's where is issue.

a®

[Former Member]

a@rs actually there is no logic involved with the VALUE field.. I just gave numbers in ascending orders for simplicity. My requirement is if I have repetitive values in NODE column then I should group the value of NODE field and respective value of VALUE field in an internal table.

Any cules how to do this?

[chaouki_akir]

Hello,

How itab is sorted ?

- For example table jtab has 3 fields D E F. D is the first one and F the last one.

-- Are itab entries sorted like that :

D  5                  <--- 1st field of structure jtab
E  4
F  7                  <--- last field of structure jtab

or like that

E  4
D 5
F 77

or something else

?

re-using the code of a®s. You can do something like that :

  
    loop at itab.
        CASE itab-node.
         WHEN 'D' OR 'E' OR 'F'.  "jtab fields
                CASE itab-node.
                   WHEN 'D'. "1st field
                          CLEAR              wa_jtab.
                          MOVE itab-value TO wa_jtab-d.
                   WHEN 'F'. "last field
                          MOVE itab-value TO wa_jtab-f.
                          APPEND wa_jtab TO jtab.
                   WHEN OTHERS.
                          ASSIGN COMPONENT itab-node OF STRUCTURE wa_jtab TO <fs>.
                          IF  sy-subrc IS INITIAL.
                               MOVE itab-value TO <fs>.
                          ENDIF.
                 ENDCASE.

         WHEN 'G' OR 'H' OR 'I'.  "ktab fields
                CASE itab-node.
                   WHEN 'G'. "1st field
                          CLEAR              wa_ktab.
                          MOVE itab-value TO wa_ktab-g.
                   WHEN 'I'. "last field
                          MOVE itab-value TO wa_ktab-i.
                          APPEND wa_ktab TO ktab.
                   WHEN OTHERS.
                          ASSIGN COMPONENT itab-node OF STRUCTURE wa_ktab TO <fs>.
                          IF  sy-subrc IS INITIAL.
                               MOVE itab-value TO <fs>.
                          ENDIF.
                 ENDCASE.

           ENDCASE.
  ENDLOOP.

Pre requisite: itab must be sorted by field position in structure of each target tables (jtab ktab ...)

Remark : if target tables have lot of fields you can use range instead of WHEN 'D' OR 'E' OR 'F'.

Cordialement,

Chaouki.

Edited by: Chaouki AKIR on Feb 19, 2009 9:59 PM

[Former Member]

Hi,

I have done this with the IF statement, As such usinf if will not efftect ur performance.

Check out the coad below it works.

TYPES: BEGIN OF t_itab,

node TYPE c,

value TYPE c,

END OF t_itab.

TYPES: BEGIN OF t_jtab,

d TYPE c,

e TYPE c,

f TYPE c,

END OF t_jtab.

DATA: i_itab TYPE STANDARD TABLE OF t_itab,

wa_itab LIKE LINE OF i_itab,

i_jtab TYPE STANDARD TABLE OF t_jtab,

wa_jtab LIKE LINE OF i_jtab,

copyy TYPE C.

LOOP AT i_itab INTO wa_itab.

IF wa_itab-node = 'D' OR wa_itab-node = 'E' OR wa_itab-node = 'F'.

IF wa_itab-node = 'D'.

wa_jtab-d = wa_itab-value.

APPEND wa_jtab TO i_jtab.

COPYY = WA_ITAB-VALUE.

CLEAR wa_jtab.

CONTINUE.

ELSEIF wa_itab-node = 'E'..

wa_jtab-e = wa_itab-value.

MODIFY I_JTAB FROM WA_JTAB

TRANSPORTING E

WHERE D = COPYY.

CLEAR wa_jtab.

ELSEIF wa_itab-node = 'F'..

wa_jtab-f = wa_itab-value.

MODIFY I_JTAB FROM WA_JTAB

TRANSPORTING F

WHERE D = COPYY.

CLEAR wa_jtab.

ENDIF.

ELSE.

CONTINUE.

ENDIF.

ENDLOOP.

I hope this will help,

Suraj

****after getting the data in the internal table.********