Loop over a table from end to beginning?

[daniel_humberg]

I want to loop over a STANDARD TABLE in the opposite direction, i.e. starting with the last entry up to the first entry.

What'S the best way to do this?

Do I have to make a copy of the table first and reverse the sorting?

Can I use LOOP in some way?

Or shall I do something like

DO n TIMES

READ TABLE t INDEX i.

ENDDO.

What if I have a HASHED TABLE (no read with index) ?

Message was edited by: Daniel Humberg

[Former Member]

Hi,

You have to copy the table & sort in descending & then loop the table.

One more thing is you cannot loop at a standard table.

[RichHeilman]

You can do something like this. LOOP does not provide this functionality.

report zrich_0001 .

data: begin of itab occurs 0,
      field1 type c,
      field2 type c,
      end of itab.

data: tablines type i.
data: lineindex type i.

itab-field1 = 'A'.
itab-field2 = '1'.
append itab.

itab-field1 = 'B'.
itab-field2 = '2'.
append itab.

itab-field1 = 'C'.
itab-field2 = '3'.
append itab.


describe table itab lines tablines.

lineindex = tablines.
do tablines times.

  read table itab index lineindex.
  write:/ itab-field1, itab-field2.
  lineindex = lineindex - 1.

enddo.

Regards,

Rich Heilman

[RichHeilman]

Also, if when building the itab, you assign an index then, then you can sort decending and a loop at it.

report zrich_0001 .

data: begin of itab occurs 0,
      index type i,
      field1 type c,
      field2 type c,
      end of itab.

data: tablines type i.
data: lineindex type i.

itab-index = '1'.
itab-field1 = 'A'.
itab-field2 = '1'.
append itab.

itab-index = '2'.
itab-field1 = 'B'.
itab-field2 = '2'.
append itab.

itab-index = '3'.
itab-field1 = 'C'.
itab-field2 = '3'.
append itab.


sort itab descending by index.

loop at itab.
  write:/ itab-field1, itab-field2.
endloop.

Regards,

Rich HEilman

[Former Member]

Here is one way.

DESCRIBE TABLE itab LINES v_numlines.

DO v_numlines TIMES.
  CLEAR itab.
  READ TABLE itab index v_numlines.
  ----- do whatever---
  v_numlines = v_numlines - 1.
ENDDO.

Srinivas

[Former Member]

Rich is lot quicker than me on the keyboard!! But I am glad that we are thinking on the same lines(.!.)

Srinivas

[Former Member]

You cannot loop at the Hashed table too.

[Washington]

Hi Daniel,

it seems strange, but i think u can do this:

describe table t lines <lines>.

while not <lines> = 0.

read table t index <lines>.

<lines> = <lines> - 1.

endwhile.

Regards,

[Former Member]

Hi,

You can select entries from Standard table into your internal table by using DESCENDING (Last entry first)by time(Field).

SELECT * FROM SE903 INTO

X_TEMP

WHERE VBELN <> ' '

AND POSNR <> ' '

ORDER BY ERDAT DESCENDING.

[RichHeilman]

I think we are talking about internal tables here....

Srinivas, Great Minds Think Alike!!!!!

Regards,

Rich HEilman

[Former Member]

Daniel,

If you can get the required state by sorting in reverse then it will be the better approach. Otherwise use your second option Do...Enddo with Read.

I don't think there is any other way to loop the internal table in opposite direction.

Thanks

Giridhar

[Clemenss]

... do this:

data:

lv_tfill type sytfill.

field-symbols:

<fs> type any.

describe table t lines lv_tfill.

while lv_tfill > 0.

read table t assigning <fs>

index lv_tfill.

write: / <fs>.

subtract 1 from lv_tfill.

endwhile.

A forward loop won't be faster.

C.

[daniel_humberg]

ok, thanks for all the answers.

Unfortunately, my table is a HASHED table, so I can't READ the table by INDEX. I can'T use SORT either, because the table not sorted by a certain field.

I am thinking of doing it like this

DATA t_copy LIKE STANARD TABLE OF t.
LOOP AT t ASSIGNING <fs>.
  INSERT <fs> into t_copy INDEX 1.
ENDLOOP.

Although it seems not very elegant..

[RichHeilman]

How about just moving your hashed table to a standard table and then reading backwards.

report zrich_0001 line-size 200 .

<b>types: begin of ttab,
       field1 type c,
       field2 type c,
       end of ttab.

data: itab type hashed table of ttab
              with unique key field1
              with header line.
data: wa like line of itab.

data: itab2 type standard table of ttab with header line.</b>

data: tablines type i.
data: lineindex type i.

start-of-selection.

  wa-field1 = 'A'.
  wa-field2 = '1'.
  insert wa into table itab.

  wa-field1 = 'B'.
  wa-field2 = '2'.
  insert wa into table itab.

  wa-field1 = 'C'.
  wa-field2 = '3'.
  insert wa into table itab.

<b>  itab2[] = itab[].</b>

  describe table itab2 lines tablines.

  lineindex = tablines.

  do tablines times.
    read table itab2 index lineindex.
    write:/ itab2-field1, itab2-field2.
    lineindex = lineindex - 1.
  enddo.

Regards,

Rich Heilman

[daniel_humberg]

yes, that'S what I do.

I just dont do a copy with itab2 = itab1, beacuse i need to sort the table in the same loop that I use for copying.