How to store results temporarily or use a subquery

[Dulanjan_Katuwawela]

So I did search for use of subqueries but most posts are from several years back, so reposting to make sure there's no feature that's come up recently.

If I have a query that results in a million IDs, how do I query against just those IDs (assuming my index contains over 100 million documents)?

In the relational world you could use store the results in a temp table and reference it in a subsequent query, or you could use a subquery which would contain the logic for the first query. What options are available at this point to do this? Or is there still no support for accomplishing this?

Thanks!!

[Dulanjan_Katuwawela]

anyone?

[warkolm (Mark Walkom)]

You need to do it externally to Elasticsearch, there is nothing it has that can do this natively.

[Mikhail_Khludnev]

probably you can do it with putting ids into the doc and referring it as a terms query
https://www.elastic.co/guide/en/elasticsearch/reference/5.5/query-dsl-terms-query.html

[warkolm (Mark Walkom)]

Not unless you get the terms first, that's the problem.

[Dulanjan_Katuwawela]

one thing we could do (which will work only if you relate data to a singular master dataset) is to reindex the results of a query, storing just the master key, establishing a parent child relationship on the created index with the master index, and use it in the form of a join (this gives you the ability to do a subquery only for a singular field, not any field though).