How to Stack Cardinality Transforms

Elastic Stack Elasticsearch
1

Hello,

This is a followup question for:

In order to delete old raw data without impacting the continuous lifetime cardinality aggregation transform, I was suggested to stack 2 transforms together. The first transform aggregated raw data into hourly summaries, then the outputs get fed into the second transform that aggregated the hourly summaries into lifetime total.

I am not sure how to set this up in practice because the output of the hourly cardinality transform is just a single integer representing how many unique values are seen in each hour, rather than some kind of data structure like HLL that can be rolled up to a higher level. Is this the right way to do this or is there a different approach?

Thanks.

2

Cardinalities can not be summarized using a stacked approach. I just read my previous post again, it sounds misleading. Sorry if I have not been clear. It should say that it does not work for cardinality.

There have been ideas about making HLL a native data type in Lucene, so it can be updated.

3

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