So, I get a bunch of log entries that look something like this (grossly simplified) example:
host1 tag - foo
host1 tag + foo
host1 tag - bar
host1 tag - something
host1 tag + something
host1 tag - evil
host1 tag + blarg
host2 tag - zoinks
I want to find the log entries that have a "- $thing" without a corresponding "+ $thing" in a 24-hour period. So for the above, I want to see "bar evil zoinks".
I can easily write a search to find the "-" entries. But when I try to exclude the ones with a corresponding "+" entry, it gets hairy. The original query already takes a while to run, and I can have thousands of matches. The obvious approach would seem to be a subsearch. But a subsearch seems like it's asking for an N-squared performance. Is there some better way to do this? I would intuitively expect that maybe a join or a selfjoin would help, but I can't figure it out. I'll keep working on this in the meantime.
Thanks!
... View more