Ok, in that case the quoted parts from the documentation are not relevant. But
especially, the requirements for stable parts of the links and those needing
replacement are not clear.
it seems in your samples, that the parentheses capture actually the parts to be
discarded, you 
Basically you need to capture the needed parts of the matched string to parens
and recombine them in the replacement pattern, e.g.:
(http://starting_part.*?of_an_url/)useles/part_.*?thereof/(useful_numbered_part_
[0-9]+)

replace with:

$1replacement_for_the_middle_part/$2

For referencing the captured groups (matched inside of parens) the
metacharacters $ folowed by a number are used in the replacement pattern - in
PSPad and apparently in your crawler tool too.
Other variants of regular expression engines use \ followed by a number here,
which also appears in your samples.

hth,
   vbr

-- 
<https://forum.pspad.com/read.php?2,71365,71383>
PSPad freeware editor https://www.pspad.com

Odpovedet emailem