Thank you, that worked. I was curious about the speed, is running a script using _source slower that doc[] ?
Totally understand a dynamic script is slower regardless of _source vs doc[]. Makes sense that having a count transformed up front during index to create a materialized value would certainly be much faster. On Thursday, January 8, 2015 at 7:04:40 PM UTC-8, Nikolas Everett wrote: > > > > On Thu, Jan 8, 2015 at 9:09 PM, Jeff Steinmetz <[email protected] > <javascript:>> wrote: > > Is there a better way to do this? >> >> Please see this gist (or even better yet, run the script locally see the >> issue). >> >> https://gist.github.com/jeffsteinmetz/2ea8329c667386c80fae >> >> You must have scripting enabled in your elasticsearch config for this to >> work. >> >> This was originally based on some comments I found here: >> >> http://stackoverflow.com/questions/17314123/search-by-size-of-object-type-field-elastic-search >> >> We would like to use a filtered query to only include documents that a >> small count of items in the list [aka array], filtering where >> values.size() < 10 >> >> "script": "doc['titles'].values.size() < 10" >> >> Turns out the values.size() actually either counts tokenized (analyzed) >> words, or if the mapping turns off analysis, it still counts incorrectly if >> there are duplicates. >> If analyze is not turned off, it counts tokenized words, not the number >> of elements in the list. >> If analyze is turned off for a given field, it improves, but duplicates >> are missed. >> >> For example, This comes back as size == 2 >> "titles": ["one", "duplicate", "duplicate"] >> This comes back as size == 3, should be 4 >> "titles": ["http://bit.ly/abc";, "http://bit.ly/abc";, "http://bit.ly/def";, >> "http://bit.ly/ghi";] >> >> Is this a bug, is there a better way, or is this just something that we >> don't understand about groovy and values.size()? >> >> >> > I think that's just the way doc[] works. Try (but don't actually deploy) > _source['titles'].size() < 10. That should do what you expect. Don't > deploy that because its too slow. Try indexing the size and filtering on > it. You can use a transform to add the size of the array as an integer > field and just filter on it using a range filter. That'd probably be the > fastest option. > > Nik > -- You received this message because you are subscribed to the Google Groups "elasticsearch" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/elasticsearch/75736948-beac-43fc-84d4-25a94456d4ca%40googlegroups.com. For more options, visit https://groups.google.com/d/optout.
