'GUI_UPLOAD': uploading of only numbers

[Former Member]

Dear Experts,

I use 'GUI_UPLOAD' for uploading of the CSV-File into my ALV and I need that whatever information this list contains, only numbers (well no letters)

will be uploaded into the grid.

My current code for uploading looks following:

CALL FUNCTION 'GUI_UPLOAD'

     EXPORTING

       filename = v_filename

       filetype = 'ASC'

     TABLES

       data_tab = ta_csv_data.

   LOOP AT ta_csv_data INTO v_string.

     IF sy-tabix = 1.

       CONTINUE.

     ENDIF.

     SPLIT v_string AT ';'

     INTO

     wa_csv-vkorg

     wa_csv-matnr

     wa_csv-preis.

     APPEND wa_csv TO ta_csv.

     CLEAR wa_csv.

Is this somehow possible?

Thank you in advance

BR, Denis

[former_member195402]

Hi,

either your field catalog has wrong definitions, or your wa_csv fields are defined numeric instead of alpha.

Regards,

Klaus

[SimoneMilesi]

Hello Denis,

how  ta_csv_data and TA_CSV are declared?

[Former Member]

Hi Simone.

ta_csv_data      TYPE STANDARD TABLE OF zst_preisupload

ta_csv      TYPE STANDARD TABLE OF zst_preisupload

Structure zst_preisupload:

UPDATE: actually I need the number only for Prices, but the Datatyp is already CURR...

[SimoneMilesi]

First of all (not related to your issue, but maybe....): why you split at ';' if you are uploading data into an internal table with the correct structure?

WA_CSV how it's declared?

My suggestion

DATA: ta_csv_data type table of string,

                  wa_csv should be a full char structure

                         vkorg type vkorg,

                          matnr type matnr,

                         preis (10)

->upload ta_csv_data as now

loop at ta_csv_data into lv_string

split as now into wa_csv

move-corresponding wa_csv to TA_CSV

append ta_csv.

endloop

[former_member195402]

Hi,

how are the WA_CSV fields defined?

WA_CSV-PREIS must be a character type for upload. After upload you may convert it to currency type.

Regards,

Klaus

[Former Member]

Thank you for fast feedbac, Simone,

well without spliting the information will be written just as a string.

I have declared WA_CSV following:

wa_csv           LIKE LINE OF ta_csv

.

[SimoneMilesi]

As stated also by Klaus, if you use split, you have to declare DESTINATION FIELDS type char.

IF you have separators, then TA_CSV_TAB should not be declare like your structure: look at it, the Zstructure got no separator!

[Former Member]

Thank Simone, Klaus for your feedbacks,

but I don't understand how declaring of the field WA_CSV-PREIS as a character, will help me to avoid the uploadning of the letters?

BR

Denis

[SimoneMilesi]

Sorry, i completly misunderstood the request.

You got your data correctly uploaded then and want to remove all non numeric characters?

[Former Member]

No problem, may be I should describe it more clear.

Yes, the data are uploaded and look good and yes I want to remove all non numeric characters

[SimoneMilesi]

From here

DATA lv_text TYPE string VALUE 'ab000cdeSDFf34534gh1ASDF234ijk5abc0'.

 WRITE : /,'Before Replace :',lv_text. REPLACE ALL OCCURRENCES OF REGEX '[[:alpha:]]' IN lv_text WITH ''. 

WRITE : /,'After Replace :',lv_text.  

Output: Before Replace : ab000cdeSDFf34534gh1ASDF234ijk5abc0  7

After Replace : 00034534123450

[[:alpha:]]

specifies any alphabet in lower or upper case.

Remember that there is no space between the single quotes in the WITH clause.

Or here

You have to do it for your char fields

[Former Member]

There can be other non-numeric characters like special character, which will not be removed when you remove all alphabets. Instead, you should replace all occurrences of regex '\D' which means anything which is not '\d', which means non-numeric characters.

[Former Member]

Thank you for the advise, Manish. I already was thoinking how to solve it in term also to avoid special characters, but when I use

REPLACE ALL OCCURRENCES OF REGEX '\D' IN v_string WITH ''

for some reason happening following:

Excel: ALV:

is there may be some possibility to replace ALL but not REGEX [digit] ?

[nikolayevstigneev]

Hi, Denis!

Have you tried to shorten the file for uploading and debugging thorougthly what's happening there?

is there may be some possibility to replace ALL but not REGEX [digit] ?

It seems to me that "\D" regex should do exactly that trick.

You can check the document for description of some more regular expressions.

[Former Member]

Your delimiter is also getting deleted when \D (non-numeric) is used. So instead of deleting all '\D', you can apply negation and delete all '[^\d;]' , assuming semicolon ; is your field separator. I haven't tested it but I think it will work fine.

In case it doesn't work, post the non-working code here.

[Former Member]

Now I kow what the problem is about.

in my string I do split by ';' and it also will be replaced by " "...

Could it be possible to make exaption for ";" somehow?

[Former Member]

Thank you Manish!

Small question: and what, when I'd like to do 2 exaptions: for ; and for , ?

It's possible?

[Former Member]

Yes, it is possible. Big brackets [ ] define a set. [abc] means it would match either a or b or c.

[^abc] would match anything other than a or b or c as negation operator ^ is used.

In your case, \d is digit and semicolon is also part of negation set. To add comma to it, use '[^\d;,]'

[Former Member]

Thank you once again!