Regular Expression: Different results when using FIND or regex classes

[mamHB]

Hi,

has anybody an idea why the FIND REGEX statement and the regex class delivers different results when searching for

^\*

in

****

I try to find every place with an asterisk as first sign of the line.

The FIND statement

DATA gv_string        TYPE        string.
DATA gv_pattern       TYPE        string.
DATA gt_match_result  TYPE        match_result_tab.
 
gv_string = '****'.
gv_pattern = '^\*'.
FIND ALL OCCURRENCES OF REGEX gv_pattern
                           IN gv_string
                      RESULTS gt_match_result.

returns one hit as expected. But the class cl_abap_regex and cl_abap_matcher returns four hits in this example:

DATA gv_string        TYPE        string.
DATA gv_pattern       TYPE        string.
DATA gt_match_result2 TYPE        match_result_tab.
DATA gx_regex         TYPE REF TO cl_abap_regex.
DATA gx_matcher       TYPE REF TO cl_abap_matcher.

gv_string = '****'.
gv_pattern = '^\*'.
TRY.
    CREATE OBJECT gx_regex
      EXPORTING
        pattern = gv_pattern.
  CATCH cx_sy_regex .
    BREAK-POINT.
    EXIT.
ENDTRY.
 
TRY.
    CREATE OBJECT gx_matcher
      EXPORTING
        regex = gx_regex
        text  = gv_string.
  CATCH cx_sy_matcher .
    BREAK-POINT.
    EXIT.
ENDTRY.
gt_match_result2 = gx_matcher->find_all( ).
BREAK-POINT.

Looks like the class doesn't consider the start of line symbol (^). Is there an error in my implementation?

Any help is appreciated.

Matthias

[mamHB]

Hi,

has anybody an idea why the FIND REGEX statement and the regex class delivers different results when searching for

^\*

in

****

I try to find every place with an asterisk as first sign of the line.

The FIND statement

DATA gv_string        TYPE        string.
DATA gv_pattern       TYPE        string.
DATA gt_match_result  TYPE        match_result_tab.
 
gv_string = '****'.
gv_pattern = '^\*'.
FIND ALL OCCURRENCES OF REGEX gv_pattern
                           IN gv_string
                      RESULTS gt_match_result.

returns one hit as expected. But the class cl_abap_regex and cl_abap_matcher returns four hits in this example:

DATA gv_string        TYPE        string.
DATA gv_pattern       TYPE        string.
DATA gt_match_result2 TYPE        match_result_tab.
DATA gx_regex         TYPE REF TO cl_abap_regex.
DATA gx_matcher       TYPE REF TO cl_abap_matcher.

gv_string = '****'.
gv_pattern = '^\*'.
TRY.
    CREATE OBJECT gx_regex
      EXPORTING
        pattern = gv_pattern.
  CATCH cx_sy_regex .
    BREAK-POINT.
    EXIT.
ENDTRY.
 
TRY.
    CREATE OBJECT gx_matcher
      EXPORTING
        regex = gx_regex
        text  = gv_string.
  CATCH cx_sy_matcher .
    BREAK-POINT.
    EXIT.
ENDTRY.
gt_match_result2 = gx_matcher->find_all( ).
BREAK-POINT.

Looks like the class doesn't consider the start of line symbol (^). Is there an error in my implementation?

Any help is appreciated.

Matthias

[Former Member]

Hi!

Not sure, but the documentation of hat (^) in ABAP's Find command says "Negation of a value set for single characters" (FIND->FIND - pattern ->Syntax of Regular Expressions ->Special Characters in Regular Expressions).

Best!

Jim

[mamHB]

Hi Jim,

thank you for your reply. But to negate a value set of single characters you have to use

[^  ]

I used just the character

^

This is defined as anchor character for the start of a line (See the documentation link you posted -> chapter 'Special characters for search strings') .

As I understand it, should the FIND REGEX statement does the same as the CL_ABAP_REGEX and CL_ABAP_MATCHER classes. Therefore, I do not understand, why both implementations deliver another result.

And I didn't find an answer for that. Does anyone else?

Matthias