Reduce number of results according to max result per field

Elastic Stack Elasticsearch
1

We have an index that has over 100 million data. it has a structure like this:

{
	// ...
	"first_name": "John",
	"last_name": "Doe",
	"company": "X Company",
	// ...
}

We want to return results let's say 3 items per company eventhough there are much more records than 3.

Here is the records:

first_name, last_name, company
John, Doe, X Company
Jane, Doe, X Company
George, Doe, X Company
William, Doe, X Company
Jack, Doe, X Company
Ellen, Doe, Y Company
Harper, Doe, Z Company
Mason, Doe, Z Company
Ella, Doe, Z Company
Scarlett, Doe, Z Company

Here is the expected query result:

first_name, last_name, company
John, Doe, X Company
Jane, Doe, X Company
George, Doe, X Company
Ellen, Doe, Y Company
Harper, Doe, Z Company
Mason, Doe, Z Company
Ella, Doe, Z Company

How can we write this kind of query? We don't want to use terms aggregation with a top hits sub-aggregation since we have 100 million records. Or do you have any idea for a performant aggregation for this kind of huge data?

2

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