Hi, I posted the following in the Kibana category, and it was suggest that it might belong here in Elasticsearch:
How many unique terms are we talking about here?
terms aggregation and the
min aggregation on a date would be one way of getting the raw data but scale is a potential issue using aggs and a single request.
In the scale I'd typically want to view (hours to weeks) , we're talking about 100s I'd say, but could be in the 10,000s at longer time periods (months and greater).
As far as I can see there's no configuration of existing aggregations that would give you the minimal set of data you are looking for (a count per time period).
The closest we can probably get is a list of all unique terms and for each the first date they appeared in the index. It would be up to your client code to organise these pairs into the required date buckets, counting the number of terms that fall into the same bucket. That could be a lot of data.
This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.