uboness f3c6108b71 introduced support for "shard_size" for terms & terms_stats facets. The "shard_size" is the number of term entries each shard will send back to the coordinating node. "shard_size" > "size" will increase the accuracy (both in terms of the counts associated with each term and the terms that will actually be returned the user) - of course, the higher "shard_size" is, the more expensive the processing becomes as bigger queues are maintained on a shard level and larger lists are streamed back from the shards. ... closes #3821		2013-10-02 22:02:00 +02:00
..
community	Add more anchor links to documentation	2013-09-30 13:13:16 -06:00
groovy-api	Add more anchor links to documentation	2013-09-30 13:13:16 -06:00
java-api	Uniquify more anchor links to fix asciidoc	2013-10-01 10:28:35 -06:00
reference	introduced support for "shard_size" for terms & terms_stats facets. The "shard_size" is the number of term entries each shard will send back to the coordinating node. "shard_size" > "size" will increase the accuracy (both in terms of the counts associated with each term and the terms that will actually be returned the user) - of course, the higher "shard_size" is, the more expensive the processing becomes as bigger queues are maintained on a shard level and larger lists are streamed back from the shards.	2013-10-02 22:02:00 +02:00
river	Add more anchor links to documentation	2013-09-30 13:13:16 -06:00