catenate_numbers is useful for joining numbers separated by hyphen, as it allows a non-segmented search term to match against segmented numbers in the document, e.g.
Document: 02 6287-1455
Search term: 0262871455 (will match with catenate_numbers)
However, sometimes numbers in the document might be like:
0262871455 0262871477 0262871499
or
0262871455,0262871477, 0262871499
However with catenate_numbers both will become one long string 026287145502628714770262871499 will which probably never match any search term.
It would be nice if we could specify the character to/to not join on with catenate_numbers, to get around this problem.