REGEXP_REPLACE issue/bug - replaces only every second occurence

Question

Hi All
I am using
Oracle Database 10g Express Edition Release 10.2.0.1.0 - Product
PL/SQL Release 10.2.0.1.0 - Production
CORE     10.2.0.1.0     Production
TNS for 32-bit Windows: Version 10.2.0.1.0 - Production
NLSRTL Version 10.2.0.1.0 - Production
I need a regular expression to rename a variable in a mathematical formula, when testing my code I discovered that if the variable was repeated right after itself only one occurence was replaced. I need to only replace whole words as some variables can be a subset of a longer variable name, and variable names can have underscores in them.
Below is an example of what happens.
Can someone please tell me why this happens? Example code is:
SELECT REGEXP_REPLACE('ab+ab+ab+ab+ab+ab+ab+abc+abc+ab', '(\W|^)ab(\W|$)' , '\1xy\2',1,0,'i') from DUAL;
The regular expression matches all 'ab' with a non alpha character or beginning line before and a non alpha or end of line at the end.
I then replace the 'ab' with 'xy' and keep the original non alpha on either side. What happens is that only every second occurence is replaced.
If I place a space after every '+' then it replaces all occurences. (Obviously this is a user field in a table of mathematical formulas so I cannot expect the user to worry about this problem.)
I believe it may be due to the fact the pointer in finding all occurence is moved incorrectly and it does not find the match immediately after.
Can someone please explain this? Or am I missing something in my example.
Thanks
Dudley

Answer 1

Is this what you require? I changed your query a bit.
SQL> SELECT REGEXP_REPLACE ('ab+ab+ab+ab+ab+ab+ab+abc+abc+ab',
  2                         '(\W|^)ab(\W|$)',
  3                         '\1xy\2',
  4                         1,
  5                         0,
  6                         'i'
  7                        )
  8    FROM DUAL;
REGEXP_REPLACE('AB+AB+AB+AB+AB+
xy+ab+xy+ab+xy+ab+xy+abc+abc+xy
1 row selected.
SQL>
SQL> SELECT REGEXP_REPLACE ('ab+ab+ab+ab+ab+ab+ab+abc+abc+ab',
  2                         '(ab)(\W|$)',
  3                         'xy\2',
  4                         1,
  5                         0,
  6                         'i'
  7                        )
  8    FROM DUAL;
REGEXP_REPLACE('AB+AB+AB+AB+AB+
xy+xy+xy+xy+xy+xy+xy+abc+abc+xy
1 row selected.
SQL> -- checking data with white spaces
SQL> SELECT REGEXP_REPLACE ('ab + ab + ab + ab + ab + ab + ab + abc + abc + ab',
  2                         '(ab)(\W|$)',
  3                         'xy\2',
  4                         1,
  5                         0,
  6                         'i'
  7                        )
  8    FROM DUAL;
REGEXP_REPLACE('AB+AB+AB+AB+AB+AB+AB+ABC+ABC+AB',
xy + xy + xy + xy + xy + xy + xy + abc + abc + xy
1 row selected.
SQL> Regards,
Jo