Help using REGEX to get the File Path from a field at search time

[mkersh]

I'm attempting to categorize my splunk usage by source - using the logfile path as a rough way to group by application. So in my search, I'm retrieving the source - and then attempting to use REGEX to just show the filepath portion of the value.

Example...

This is my search query...

index=_internal source=license_usage type=Usage | eval MB=round((b/1024)/1024,2) | stats sum(MB) AS usage(mb) by h,s | sort 0 -usage(mb) | rename s as sourceDir | rename h as host

This is a sample of the values being returned for s

/cgu/prd/data/applogs/technical/uapesb02N1.PrdCguIccWesbAppM2.ServiceActivityEvent.log
/pi/prd/data/scv/CRODS_R3/logs/crods_r3.log

/cgu/prd/http/PrdCgu2Http/logs/access_log.2012-03-30

I've tried to use REX to remove the filename and just show the filepath portion using this...

| rex field=s mode=sed "s/(^.*)\(?!\)//g" |

but so far nothing seems to be working. Please help 🙂

[Damien_Dallimor · Accepted solution]

Does this do the trick ?

... | rex field=s "(?<filepath>(/\w+)+)/" | table filepath

[mkersh]

That worked perfectly. Thanks!

[Damien_Dallimor]

updated the original answer and verified that it works in Splunk.

[mkersh]

Thanks - that's closer... It is returning just the last path... so /pi/prd/data/scv/CRODS_R3/logs/crods_r3.log becomes /logs