Regular Expressions
Replacing
The REPLACE
statement can work with regular expressions directly:
DATA(lv_test) = 'The quick brown fox'.
REPLACE ALL OCCURRENCES OF REGEX '\wo' IN lv_test WITH 'XX'.
The variable lv_test
will evaluate to The quick bXXwn XXx
.
Searching
The FIND
statement can work with regular expressions directly:
DATA(lv_test) = 'The quick brown fox'.
FIND REGEX '..ck' IN lv_test.
" sy-subrc == 0
FIND REGEX 'a[sdf]g' IN lv_test.
" sy-subrc == 4
Object-Oriented Regular Expressions
For more advanced regex operations it’s best to use CL_ABAP_REGEX
and its related classes.
DATA: lv_test TYPE string,
lo_regex TYPE REF TO cl_abap_regex.
lv_test = 'The quick brown fox'.
CREATE OBJECT lo_regex
EXPORTING
pattern = 'q(...)\w'.
DATA(lo_matcher) = lo_regex->create_matcher( text = lv_test ).
WRITE: / lo_matcher->find_next( ). " X
WRITE: / lo_matcher->get_submatch( 1 ). " uic
WRITE: / lo_matcher->get_offset( ). " 4
Evaluating Regular Expressions with a Predicate Function
The predicate function matches
can be used to evaluate strings on the fly without use of any object declarations.
IF matches( val = 'Not a hex string'
regex = '[0-9a-f]*' ).
cl_demo_output=>display( 'This will not display' ).
ELSEIF matches( val = '6c6f7665'
regex = '[0-9a-f]*' ).
cl_demo_output=>display( 'This will display' ).
ENDIF.
Getting SubMatches Using OO-Regular Expressions
By using the method GET_SUBMATCH
of class CL_ABAP_MATCHER
, we can get the data in the groups/subgroups.
Goal: get the token to the right of the keyword ‘Type’.
DATA: lv_pattern TYPE string VALUE 'type\s+(\w+)',
lv_test TYPE string VALUE 'data lwa type mara'.
CREATE OBJECT ref_regex
EXPORTING
pattern = lv_pattern
ignore_case = c_true.
ref_regex->create_matcher(
EXPORTING
text = lv_test
RECEIVING
matcher = ref_matcher
).
ref_matcher->get_submatch(
EXPORTING
index = 0
RECEIVING
submatch = lv_smatch.
The resulting variable lv_smatch
contains the value MARA
.