This entry is to help me next time I need to do the same thing.

I frequently run XPath in oXygen to find stuff. For example, I just wanted to find all the non-conformant xml:id attributes in a file. So, I used this XPath expression to find them:

//person[matches(@xml:id,'[a-z]{5}')

and it responded with output like this:

System ID: /path/to/my/file/FILENAME.xml
Description: xml:id="feefifofum_3"
XPath location: /TEI[1]/text[1]/body[1]/div[1]/listPerson[2]/person[166]
Start location: 1703:21
End location: 1708:30

ad nauseam

I'm only interested in the actual xml:id value (feefifofum_3) so I ran this regex:

^System.*\nDescription: xml:id="(.*)?".*\n.*\n(^Start.*\n)?(^End.*\n)?

and replaced the match with the first group \1 (or $1 in oXygen)