Ignore the bit about high cardinality.

IMO, the use of fielddata for date_historgram is counter intuitive, since 
the buckets made a ranges as defined by the grain, and therefore the 
buckets are pre-defined in a sense which means the inverted index should 
itself be sufficient for building the buckets, fielddata is not required.

Is this how a date_histogram actually behaves under the covers ?

On Tuesday, March 10, 2015 at 1:43:13 PM UTC-7, Anant Aneja wrote:
>
> I was reading up on when ES builds fielddata : 
> https://www.elastic.co/guide/en/elasticsearch/guide/current/fielddata.html
>
> A footnote on this page reads 'It is required for any operation that needs 
> to look up the value contained in a specific document'.
>
> Would this also be true when building buckets in a date_histogram 
> aggregation ? Date fields are bound to have high cardinalilty so I would 
> imagine that if populating fielddata would be counterproductive.
>
> In what cases while building aggregations does the above mentioned pattern 
> not apply ?
>
>
>

-- 
You received this message because you are subscribed to the Google Groups 
"elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To view this discussion on the web visit 
https://groups.google.com/d/msgid/elasticsearch/fdd817e6-7ac2-49be-862b-947492da2dda%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Reply via email to