Remove timestamps using regex

[oliverpeloton23]

Hi Splunk Community,

It's been a while since I've last used Splunk and regex, and now I'm struggling with both 🙂

Fields that I need to use ("resourceId") contain two user IDs and timestamps (e.g., "owner-10785-user-3801-key-1637099215"). I'm looking to keep the IDs and remove timestamps (basically everything after "owner-19803-user-8925-").

I came up with this clumsy thing:

index=main | eval resourceId1=replace (resourceId, "user-(?<user_id>\d+)", "") | eval resourceId2=replace (resourceId1, "owner-(?<owner_id>\d+)", "") | table resourceId2

It kind of works, the only problem is that it gives me the opposite result - it removes all the IDs leaving the timestamps, like this:

resourceId2
--key-1637100297
--1637100120.0929909
--key-1637100118

But I need the opposite. Can anyone please help?

[ITWhisperer · Accepted solution]

| rex field=resourceId "owner-(?<owner>\d+)-user-(?<user>\d+)"
| eval resourceId2="owner-".owner."-user-".user

[oliverpeloton23]

Thank you!