Why do aggregation queries on keyword field produce fieldata in JVM HEAP?

MarkDai · July 10, 2018, 11:21am

Hi all,

Does anyone know why making aggregations on keyword fields of an index produce fieldata cache in JVM HEAP?
Per the docs on elastic official site, most fields except text will use doc_values to support aggregations. So it should not produce fielddata cache in JVM heap when making aggregations on keyword field ?

https://www.elastic.co/guide/en/elasticsearch/reference/6.3/fielddata.html

Thanks!

Igor_Motov · July 13, 2018, 4:32pm

Which version of elasticsearch are you using? Can you show how your query and mapping look like? Can you check which fields are getting the fielddata associated with them by running:

GET /_nodes/stats/indices/fielddata?fields=*

system · August 10, 2018, 4:32pm

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.

Topic		Replies	Views
When can keyword field with doc values enabled consume fielddata cache? Elasticsearch	3	605	August 10, 2018
Term aggregation on frozen index cause field data cache increase continuously Elasticsearch	2	305	April 28, 2020
ElasticSearch term Aggregation on text fields Elasticsearch	1	405	June 25, 2020
Resource upper bounds on keyword sorting Elasticsearch	1	325	January 6, 2021
Spike in Fielddata memory usage Elasticsearch	1	443	June 25, 2019

Why do aggregation queries on keyword field produce fieldata in JVM HEAP?

Related topics