topic Re: write a file by using open dataset in Application Development and Automation Discussions

write a file by using open dataset

Former Member — Wed, 13 Jul 2005 10:23:12 GMT

Hi all,

I have to write a log from an internal table ITAB to a file in this format:

COL1 | COL2 | COL3 | COL4

123 11 ABC 3

What should I do?

Re: write a file by using open dataset

jayanthi_jayaraman — Wed, 13 Jul 2005 10:28:27 GMT

Hi,

Check this link.You can find sample code.

http://www.sts.tu-harburg.de/teaching/sap_r3/ABAP4/open_dat.htm

Re: write a file by using open dataset

Former Member — Wed, 13 Jul 2005 10:30:29 GMT

Hi,

you can try something like this -



DATA HEADER(72).

CONCATENATE 'COL1' 
            'COL2'  
            'COL3'  
            'COL4' 
       INTO HEADER 
  SEPARATED BY SPACE.

LOOP AT ITAB.
  AT FIRST.
    TRANSFER HEADER TO FILENAME.
  ENDAT.
  CONCATENATE ITAB-COL1
              ITAB-COL2
              ITAB-COL3
              ITAB-COL4
         INTO FILE_LINE
    SEPARATED BY '|'.
  TRANSFER FILE_LINE TO FILENAME.
ENDLOOP.

Regards,

Anand Mandalika.

Re: write a file by using open dataset

Vinod_Chandran — Wed, 13 Jul 2005 10:33:47 GMT

If you want to write to the PC then use GUI_DOWNLOAD

To the application server use OPEN, TRANSER and CLOSE

DATA: file(20) TYPE C value '/usr/test.dat'.

OPEN DATASET file IN TEXT MODE ENCODING DEFAULT FOR OUTPUT.

IF SY-SUBRC = 0.

LOOP AT ITAB.

TRANSFER ITAB TO file.

ENDLOOP.

ENDDO.

CLOSE DATASET file.

Re: write a file by using open dataset

Former Member — Wed, 13 Jul 2005 10:34:46 GMT

Download internal table to Application server file(Unix)

DATA: e_file like rlgrap-filename value '/usr/sap/tmp/file.txt'.

open dataset e_file for output in text mode.

lOOP AT it_datatab......

transfer it_datatab to e_file.

ENDLOOP.

close dataset e_file.

Hope this helps.

Thanks & Regards,

Judith.

Re: write a file by using open dataset

andreas_mann3 — Wed, 13 Jul 2005 10:40:09 GMT

Hi Macy,

or try to do it dyn. with assign component:

(for big structures like bseg)

1) heading:

-get fieldcatalog of your itab

 loop at fcat.
   CONCATENATE 'ITAB-' fcat-name INTO hfield.

    ASSIGN (hfield) TO <fs>.
    CONCATENATE str  <fs> INTO str SEPARATED BY space.

 endloop.
 transfer str to file.

2) data

  LOOP AT itab.
    clear str.
    DO.
      ASSIGN COMPONENT sy-index OF STRUCTURE itab TO <fw>.
      IF sy-subrc <> 0.
        EXIT.
      ENDIF.
   CONCATENATE str <fw> INTO str SEPARATED BY space.
  ENDDO.

     transfer str to file.

  ENDLOOP.

regards Andreas

Message was edited by: Andreas Mann

Re: write a file by using open dataset

Former Member — Wed, 13 Jul 2005 10:42:53 GMT

Hi ,

Open the file using the open dataset statement in output mode.

Then transfer the contents of the internal table to the file.

sample code:

open dataset w_dsn for output in text mode encoding default.

LOOP AT ITAB.

IF SY-TABIX = 1.

ITAB5_OUT-MATNR = 'INPUT MATERIAL #..'.

ITAB5_OUT-WERKS = 'POM.'.

ITAB5_OUT-PHWERKS = 'PHPL'.

TRANSFER ITAB5_OUT TO DSN.

ENDIF.

MOVE-CORRESPONDING ITAB TO ITAB5_OUT.

TRANSFER ITAB5_OUT TO DSN.

ENDLOOP.

CLOSE DATASET DSN.

here itab5_out is the internal table that has the headers.Itab is the internal table that contans the actual data.

Hope this helps you.

Regards,

Venkatalakshmi

Re: write a file by using open dataset

Former Member — Thu, 14 Jul 2005 01:44:43 GMT

Hi Anand Mandalika,

The sample codes are very useful to me. Thanks a lot.

But it can't insert a break when I write the actual content after the header line, i.e. the header and the content line is on the same line in the file.

How to add a break for this?

Thanks.

Re: write a file by using open dataset

Former Member — Thu, 14 Jul 2005 02:35:25 GMT

Hi Macy,

This is something that I have not noticed before. But I've found the solution as well. Here's what you need to do -

DATA HEADER(72).
 
CONCATENATE 'COL1' 
            'COL2'  
            'COL3'  
            'COL4' 
       INTO HEADER 
  SEPARATED BY SPACE.
 
LOOP AT ITAB.
  AT FIRST.
<b>    open dataset filename for output in text mode.</b>
    TRANSFER HEADER TO FILENAME.
  ENDAT.
  CONCATENATE ITAB-COL1
              ITAB-COL2
              ITAB-COL3
              ITAB-COL4
         INTO FILE_LINE
    SEPARATED BY '|'.
  TRANSFER FILE_LINE TO FILENAME.
ENDLOOP.

The only difference to the earlier code, as you can see , is the OPEN DATASET statement.

Regards,

Anand Mandalika.

Re: write a file by using open dataset

Former Member — Thu, 14 Jul 2005 02:48:26 GMT

Hi Macy,

There's one more thing that you need to know here. If tany of the fields in a record of the internal table is

initial, then you will get some alignment problems for the separator (in this case, '|'). For example, if your

internal table has 3 columns and the following entries,

01	ONE	SUNDAY	
02	TWO	MONDAY
03		TUESDAY
04	FOUR	WEDNESDAY
05	FIVE	THURSDAY
06	SIX	FRIDAY
07		SATURDAY

then your output in the file is going to look like this -

01|ONE|SUNDAY	
02|TWO|MONDAY
03||TUESDAY
04|FOUR|WEDNESDAY
05|FIVE|THURSDAY
06|SIX|FRIDAY
07||SATURDAY

Clearly, this is undesirable. As you can see above, this problem occurs even when the length of the content of the

fields in the internal table is uneven.

So I suggest that you do something like this -

DATA: BEGIN OF ITAB OCCURS 0,
        CARRID LIKE SFLIGHT-CARRID,
        SEPARATOR1,
        CONNID LIKE SFLIGHT-CONNID,
        SEPARATOR2,
        FLDATE LIKE SFLIGHT-FLDATE,
      END OF ITAB.

DATA : HEADER(20) VALUE 'SPFLI DETAILS',
       FILENAME(60) VALUE <give the name of the AppServer file here>,
       FILE_LINE(40).


SELECT CARRID
       CONNID
       FLDATE
  FROM SFLIGHT
  INTO CORRESPONDING FIELDS OF TABLE ITAB.

ITAB-SEPARATOR1 = '|'.
ITAB-SEPARATOR2 = '|'.

MODIFY ITAB TRANSPORTING SEPARATOR1 
                         SEPARATOR2 
                   WHERE SEPARATOR1 EQ SPACE.


LOOP AT ITAB.
  AT FIRST.
    OPEN DATASET FILENAME FOR OUTPUT IN TEXT MODE.
    TRANSFER HEADER TO FILENAME.
  ENDAT.
  TRANSFER ITAB TO FILENAME.
ENDLOOP.

Hope that make the point clear. If yes, I wouldn't mind more points :-). Otherwise, please do get back. And I'll

try to help you understand.

Regards,

Anand Mandalika.