Distinct Events Days in Index

[fredclown]

I'm trying to write an efficient search to find out the distinct days of events that I have in an index. Basically, I want to be able to see if I am missing data for any days. I could pipe the date in mm/dd/yyyy format to a table and then use dedup, but that is SLOW and my hunch is there is a better and faster way to do it.

[somesoni2 · Accepted solution]

Try this one. decent performance.

index=yourindex| stats first(_time) as _time by date_mday | convert timeformat="%Y/%m/%d" mktime(_time) as Date

[fredclown]

Yeah, that is MUCH faster than my previous attempt. Thanks sir.

[lukejadamec]

You could create a summary index, and then back fill it with data in the background. It would take a while to backfill, but it does that in the background. Once it is done, it would be very fast to search, and re-search. Here is the link to the summary index page.

http://docs.splunk.com/Documentation/Splunk/6.0.1/Knowledge/Usesummaryindexing