topic List almost-duplicates (in a variable?) in Splunk Search

List almost-duplicates (in a variable?)

djm229 — Mon, 25 Jan 2021 18:44:48 GMT

Hello. I have a search that results in, amongst other things, fields that are ALMOST duplicates. Example:

Bob: Task incomplete.

Steve: Task Incomplete.

George: Task Incomplete.

Fred: Task Complete.

Spock: Task Complete.

Is there a way to generate a table which results in:

| Task Status | Team Player Responsible |

| Task Complete | Fred, Spock |

| Task Incomplete | Bob, Steve, George |

For each instance of "Task Complete" and "Task Incomplete", I want to know which Persons to tag in a simple table without listing every user in a giant list.

I've looked into dedup, uniq and sub-searches but haven't figured out how to list the non-duplicated parts next to a single instance of the duplicated part.

[Insert deity], I hope that made sense. Help?

Re: List almost-duplicates (in a variable?)

scelikok — Mon, 25 Jan 2021 20:40:05 GMT

Hi @djm229,

I assumed your usernames and task status is in the same field. If not rex command is unnecessary;

Please try below by updating your <your_field> with field that contains your sample values;

| rex field=<your_field> "(?<user>[^:]+):\s(?<task_status>.+)" | stats values(user) by task_status

Re: List almost-duplicates (in a variable?)

djm229 — Tue, 26 Jan 2021 19:43:09 GMT

Thanks, I'll have to modify the regex to make it (and the rest of my query) fit together. If it works, I'll tag as "the Solution". Very much appreciate it - even if it doesn't work.