Typically storing source is the right way to go but if you frequently
highlight or return small fields (< 100KB) on documents that contain large
fields (> 2MB) you'll start to see decoding the source be a performance hit.
On the other hand this is something that the Elasticsearch team is talking
about working on so I expect that problem will disappear in a while. Its a
bit of a complex issue but its a recognized one.
@nik9000 - Looks like I'm running into the issue you explained. My _source is big in size and I'm using nested mapping. I've bunch of fields but I'm interested to retrieve only a few. One of the field to be retrieved is nested field. I tried setting store to true for nested field but looks like I'm running into this issue: https://github.com/elastic/elasticsearch/issues/5245
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.