DataMystic

I'm moving and cleaning up a website (about 40k pages) I took over maintence of that has several problems. Some people on a forum recommended I check out text pipe for some of the things I needed to do. So far many of the problems I have appear to be correctable with text pipe, but I don't see how to do one of the really major cleanup tasks I have.

The site has tons of internal page links that currently appear like:
../www.website.com/section/Reference/Entry ... x.htm#e955
../www.website.com/section/Reference/Entry ... x.htm#e968
../www.website.com/section/Reference/Entry ... .htm#e1425
../www.website.com/Definition/a-c/655/index.htm#d655
../www.website.com/Definition/a-c/2265/index.htm#d2265

the ../www.website.com/ and index.htm part of the URLs is constant but the rest is variable and there's several thousand different variations of the code that falls between them. What I want to do is drop everything but the /####/index.htm#xxxx to make simple relative links. So the examples above would eventually become something (I can easily adjust the paths and such after the main URL cleanup is done) like:

../955/index.htm#e955
../968/index.htm#e968
../1425/index.htm#e1425
../655/index.htm#d655
../2265/index.htm#d2265

But I don't see a way for text pipe to do it. I can get it to match the full URLs easily enough with
../www.website.com/[ longest 1 to 40 letters or digits or <!"#$%&'()*+,-./\:;=?@[]^_`{}~|> ]/[ longest 1 to 4 digits ]/index.htm[ longest 1 to 5 digits or letters or <!"#$%&'()*+,-./\:;=?@[]^_`{}~|> ]
But I don't see any way to get it to remove the preceeding code and leave the part I want to keep nor anyway to match the offending text and just remove it without the part I want to keep.

Is it possible to to do this with Text Pipe?

You need to use the EasyPattern capture() function for the bits you want to keep, then use variables $1.. in the replacement expression:

../www.website.com/[ longest 1 to 40 letters or digits or <!"#$%&'()*+,-./\:;=?@[]^_`{}~|> ]/[ capture(longest 1 to 4 digits) ]/index.htm[ capture(longest 1 to 5 digits or letters or <!"#$%&'()*+,-./\:;=?@[]^_`{}~|>) ]

Replace with

$1\$2\