0

I have an index with 10M documents with nearly 1M unique names and 20 shards, when I use terms aggregation it produces an approximate aggregation result as an CURL request, but when I try to replicate the approximation in REST high level client in Java, it produces accurate aggregation results.

Is there any way to replicate the approximate aggregation results using the REST hlc? Thanks.

achieved in Elasticsearch-head plugin - shows 7 instead of 13, actual count in kibana

nick
  • 1
  • 3
  • what do you mean by accurate and approximate? – warkolm Sep 29 '21 at 07:42
  • Approximate count of occurrence of an word in an document, when using terms aggregation, I'll add images to explain. – nick Sep 29 '21 at 07:57
  • you're running an aggregation in one screenshot and a terms filter in another though, so it'd be better to use the same query in both approaches – warkolm Sep 30 '21 at 00:40

0 Answers0