REGEX - find zip code in text

[Former Member]

Hi experts!

I have to find the zip code (for germany) in a given text. I found 5 digits in my text, but my problem is to ignore numbers which consists of more than 5 digits!

My first try works for all cases but not for the last one.

FIND FIRST OCCURRENCE OF REGEX '([0-9]{5})' IN ld_string SUBMATCHES ld_plz.

D-12345 Mainz -> should match 12345

D 12345 Mainz -> should match 12345

12345 Mainz -> should match 12345

12345Mainz -> should match 12345

Mainz D-12345 -> should match 12345

D-123 45 Mainz -> error because of the space between the numbers

D-12333345 Mainz -> error because only 5 digits are valid for a germany zip code; my REGEX does not work!

Thanks a lot!

Regards,

Florian

[Former Member]

D-123 45 Mainz -> error because of the space between the numbers

find ` ` in LD_STRING. " ` `  is back quotes (quote beside 1 in keyboard) 
if sy-subrc = 0
  message 'error, space not allowed'.
endif

and for this

D-12333345 Mainz -> error because only 5 digits are valid for a germany zip code;

i guess your code is correct. but you can try removing the ( )s

find regex '[0-9]{5}' in ld_string.

[Clemenss]

Where's the question?

I think the introduction of REGEX in ABAP is revolutionray!

Regards,

Clemens

[former_member156446]

Hi Florian check this :

FIND FIRST OCCURRENCE OF REGEX '^([0-9]{5})$' IN ld_string SUBMATCHES ld_plz.

[Former Member]

Hi J@Y!

That's what I tried first. But unfortunately it does not seem to work. I tested it with DEMO_REGEX_TOY and it does not match in any case!? I wonder why!? I think that ^ and $ stands the start and the end of the hole content in the variable LD_STRING. It only matches 12345.

Regards,

Florian

[Clemenss]

Hi Florian,

curiosity persists.

After some playing around I reduced the pattern using Placeholder for any single digit \d. Then I noticed that the 5-digit-sequence will also match 5 digits out of 6, so i used \D Placeholder for any character other than a digit Then I don't know how to recognize (optional) line start or end as alternative to non-digit, so I just enclose the string to be checked into spaces - please suggest a more elegant solution.

My test form

FORM regex .
  DATA:
    lv_subm   type string,
    lt_string TYPE TABLE OF string.
  FIELD-SYMBOLS:
    <string>  TYPE string.
  APPEND:
    'D-12345 Mainz' TO lt_string,
    'D 12345 Mainz' TO lt_string,
    '12345 Mainz'   TO lt_string,
    '12345Mainz'    TO lt_string,
    '123456Mainz'    TO lt_string,
    '123 45Mainz'    TO lt_string,
    'Mainz D-12345' TO lt_string.
  LOOP AT lt_string ASSIGNING <string>.
    clear:
      lv_subm.
    CONCATENATE ` ` <string> ` ` into <string>.
    FIND REGEX '\D(\d{5})\D' IN <string> SUBMATCHES lv_subm.
    WRITE: / <string>, 20 'matches', 30 lv_subm,40 'SY-SUBRC=', sy-subrc.
  ENDLOOP.
ENDFORM.                    " REGEX

creates this output:

 D-12345 Mainz     matches   12345     SY-SUBRC=     0
 D 12345 Mainz     matches   12345     SY-SUBRC=     0
 12345 Mainz       matches   12345     SY-SUBRC=     0
 12345Mainz        matches   12345     SY-SUBRC=     0
 123456Mainz       matches             SY-SUBRC=     4
 123 45Mainz       matches             SY-SUBRC=     4
 Mainz D-12345     matches   12345     SY-SUBRC=     0

As I do not fully understand the meaning of FIRST OCCURRENCE, I just removed it.

Regards,

Clemens

[rainer_hbenthal]

Do not try to put everything in one regex, it makes them non performant, hard to read and hart to maintain.

As a first approach i would add word boundaries:

FIND FIRST OCCURRENCE OF REGEX '(\<[0-9]{5}\>)' IN ld_string SUBMATCHES ld_plz.

This fixes most of your examples but 12345Mainz not because the numbers do not have a word boundary.

If the regex above wouldfail, i would try another regex which fits the last remaining exmaple, maybe in alist with user approval.