My understanding of the documentation (and my experiments) is that the inner
keeps only events that match both searches, left
keeps matches as well as items that were in the parent search but not in the child search, and outer
keeps ONLY events that did not match from both searches. Unfortunately, I can't figure out how to keep items from the subsearch.
The search that I have set as the parent search returns more than 50k events and therefore can't be the subsearch.
Any ideas how to use the join
command where the search keeps any events from the subsearch that do not have matches with the parent?
You might be able to do it without join (using stats). Could you provide your two searches?
Check this link.. how to join with out common fields
You might be able to do it without join (using stats). Could you provide your two searches?
Thanks. That is exactly how I interpreted your tip. Ideally I get them on the same initial search but at the least I think I'll be all set using stats. Thanks again!
Just for reference (see 2nd answer) http://answers.splunk.com/answers/81741/full-outer-join
Yes! Stats! I think that's the way to go. Then I can use something simple like a | fields just the stuff to keep | stats latest(*) as * by unique1 unique2
Good idea!
I can't really share the actual search - they are derived from a massive amount of macros and I'm concerned they exposes a little more about the business than I'm comfortable with.
The stats is a great idea - exactly what I was looking for - a way to approach this from another direction. Thanks again.