Significant field aggregation ElasticSearch

Elastic Stack Elasticsearch
1

Is there something similar to ElasticSearch's significant_text aggregation, but for significant_fields?

Example: based on the foreground of Json documents, there is an uncommonly common occurrence of the field "RAM capacity" when doing a query on computers. Is there an aggregation or Json query type that will return those uncommonly common field names (RAM capacity, CPU type, HDD capacity, etc)?

Thanks!

2

Nothing out of the box.
A couple of options spring to mind:

  1. Do-it-yourself in the client-side by performing the calculation on foreground/background numbers obtained from a combination of the global agg (for background stats) and the filters agg with exists queries for the fields of interest
  2. Re-index the content with a fieldnames array of the type keyword that lists the fields available on a doc. Using significant_terms agg to compute the significance of values.
3

Thanks Mark!

1 Like
4

FYI the field exists query and the adjacency matrix agg can be used together to show how doc types and fields relate: https://youtu.be/JzHRuJOCnR0

5

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