How to write ABAP code to split a sentence into Word by Word and store them

[Former Member]

Hi all,

I have to split a sentence into word by word and store them into separate columns. This sentence is nothing but a text description (TXTMD) from table TXXXX. The description will have 3 or more than 3 words and LAST word will be <b>always with %</b>'. The following is the sample input data:

KEY(ORD43) Description(TXTMD)

'AAAA' '1234 ABCD COMPANY USA 80%'

'BBBB' '4745 XYZ INC CANADA ABCD 70%'

For the first record:

'1234' should stored in to <b>Field 1</b>, 'ABCD COMPANY USA' into <b>Field 2</b> and

'80%' into <b>Filed 3</b>.

For the second record:

'4745' should stored in to <b>Field 1</b>, 'XYZ INC CANADA ABCD' into <b>Field 2</b> and 70%' into <b>Field 3</b>.

The first word into Field 1, the last word with % into Field 3 and all middle words into Field 2. The number of words in the sentence could be 3 or more than 3.

Could you please help in writing the ABAP for this requirement ????

Thanks in advance.

Regards,

Venkat.

[Former Member]

Use the SPLIT command.

Rob

[Former Member]

Hi Rob,

The first word into Field 1, the last word with % into Field 3 and all middle words into Field 2. The number of words in the sentence could be 3 or more than 3.

Thanks,

Vena.

[Former Member]

It was just meant to get you going. You can easily add the logic to re-combine and get the results you need.

Rob

[Former Member]

hi ...

define an internal table..

<b>data :begin of itab occurs 0,

field(15),

end of itab.

data w_line type i.

split TXMTD at space into table itab.

describe table itab lines w_line.

loop at itab.

if sy-tabix eq 1.

var1 = itab-field.

continue.

endif.

if sy-tabix eq w_line.

var3 = itab-field.

continue.

endif.

concatenate var2 itab-field into var2 separated by space.

endloop.</b>

Please do remember to close the thread when ur problem is solved !!!

Regards,

Sai Ramesh

[Former Member]

<b>split txtmd at space into variable1 variable2.</b>.........

Regards,

Sai Ramesh

[Former Member]

Try something like this.

DATA: BEGIN OF itab OCCURS 0,
        key(4),
        descr(40).
DATA: END OF itab.

DATA: BEGIN OF itab_words OCCURS 0,
        word(20).
DATA: END OF itab_words.

DATA: BEGIN OF itab_final OCCURS 0,
        field1(4),
        field2(30),
        field3(4).
DATA: END OF itab_final.

itab-key   = 'AAAA'.
itab-descr =  '1234 ABCD COMPANY USA 80%'.
APPEND itab.
CLEAR itab.

itab-key   = 'BBBB'.
itab-descr = '4745 XYZ INC CANADA ABCD 70%'.
APPEND itab.
CLEAR itab.

LOOP AT itab.
  SPLIT itab-descr AT space INTO TABLE itab_words.
  DO.
    READ TABLE itab_words INDEX sy-index.
    IF sy-subrc <> 0.
      EXIT.
    ENDIF.
    IF sy-index = 1.
      itab_final-field1 = itab_words-word.
      CONTINUE.
    ENDIF.
    IF itab_words-word CA '%'.
      itab_final-field3 = itab_words-word.
      APPEND itab_final.
      CLEAR itab_final.
    ELSE.
      CONCATENATE itab_final-field2
                  itab_words-word
             INTO itab_final-field2 SEPARATED BY space.
    ENDIF.
  ENDDO.
ENDLOOP.

LOOP AT itab_final.
  WRITE:/ itab_final-field1,
          itab_final-field2,
          itab_final-field3.
ENDLOOP.

[Former Member]

Hi Srinivas,

Thank you very much for the detailed response.

Sorry, for not giving the complete details.

It has to read the table TXXXX in a loop with Key Field KEYFIG and split the words for each record. I can not hardcode in the program.

Thanks again.

Regards,

Venkat.

[Former Member]

I hardcoded the records only as an example, but you can fill the same itab with records from a database table and then do the rest of it.

[Former Member]

Hi,

You the below code :

report ztest.

data : begin of itab1 occurs 0,
        txt100(100) type c,
       end of itab1.


data : begin of itab occurs 0,
        bukrs like bsis-bukrs,
        txt50 like skaT-txt50,
        per(4) type c,
      end of itab.

DATA : LINE TYPE I,
      LEN TYPE I,
      TEMP TYPE I,
      TEMP1 TYPE I.


ITAB1-txt100 =  '1234 ABCD COMPANY USA 80%'.
APPEND ITAB1.
CLEAR ITAB1.

ITAB1-txt100 = '4745 XYZ INC CANADA ABCD 70%'.
APPEND ITAB1.
CLEAR ITAB1.


LOOP AT ITAB1.

  CALL FUNCTION 'STRING_LENGTH'
    EXPORTING
      string       = ITAB1-txt100
   IMPORTING
     LENGTH        = LEN.

TEMP = len.

TEMP = LEN - 7.

ITAB-BUKRS = ITAB1-TXT100.
ITAB-TXT50 = ITAB1-TXT100+4(TEMP).

TEMP1 = TEMP + 4.
ITAB-PER = ITAB1-TXT100+TEMP1(3).

APPEND ITAB.
CLEAR ITAB.

ENDLOOP

loop at itab.
write : /  ------
endloop.

Thanks,

Sriram Ponna.

[Former Member]

REPORT ztest1111

NO STANDARD PAGE HEADING LINE-SIZE 255.

PARAMETERS: x TYPE string.

DATA: y TYPE string,

f1 TYPE string,

f2 TYPE string,

f3 TYPE string,

temp TYPE i,

temp1 TYPE i.

MOVE x TO y.

SPLIT y AT space INTO f1 f2.

temp = STRLEN( f2 ).

temp1 = temp - 3.

f3 = f2+temp1(3).

WRITE: /'field1',f1.

WRITE: /'field2',f2.

WRITE: /'field3',f3.

[Former Member]

Hi Venkat,

This is a simple approach assuming the field length of the last field is fixed to 3.

PARAMETERS: x TYPE string.

DATA: y TYPE string,

f1 TYPE string,

f2 TYPE string,

f3 TYPE string,

temp TYPE i,

temp1 TYPE i.

MOVE x TO y.

SPLIT y AT space INTO f1 f2.

temp = STRLEN( f2 ).

temp1 = temp - 3.

f3 = f2+temp1(3).

WRITE: /'field1',f1.

WRITE: /'field2',f2.

WRITE: /'field3',f3.

[former_member195698]

Another way to perform your task....

wf_text = '1234 ABCD COMPANY USA 80%'.

split wf_text into wf_field1 wf_temp at ' '. (This is put the first word into field1 and rest in wf_temp)

Call the function STRING_REVERSE for wf_temp.

split the wf_temp into wf_field3 wf_field2 at ' '. (This will put reversed text into field2 and field3)

Call the Function STRING_REVERSE for wf_field2 and wf_field3.

Thanks & Regards,

Abhishek

[Former Member]

Easy way if you know the length of the field.

In my example I assumed my field has 12 places.

DATA: obs TYPE string,

       bkp TYPE char12.

obs = '12345678901234567890123456789012345678901234567890'.

DO.

   MOVE obs TO bkp.

   SHIFT obs BY 12 PLACES LEFT.

   write: / bkp.   "Do what you want with bkp variable

   IF obs IS INITIAL.

     EXIT.

   ENDIF.

ENDDO.

[nabheetscn]

Celito its a 6 year old thread:)

[Former Member]

I know, but I needed the solution, didn't found.

This topic is the first result from Google, so others may need too.