You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
String terms aggregations today work by accumulating counts into a hash table that stores the term values as keys (just like facets when provided map as an execution_hint).
Similarly to facets, we should also have an execution mode that allows to build the buckets based on string ordinals. This proved to be much faster than the map execution mode for facets, so hopefully it should help speed-up string terms aggregations as well.
The text was updated successfully, but these errors were encountered:
When the ValuesSource has ordinals, terms ordinals are used as a cache key to
bucket ordinals. This can make terms aggregations on String terms significantly
faster.
Closeelastic#4350
String terms aggregations today work by accumulating counts into a hash table that stores the term values as keys (just like facets when provided
map
as anexecution_hint
).Similarly to facets, we should also have an execution mode that allows to build the buckets based on string ordinals. This proved to be much faster than the
map
execution mode for facets, so hopefully it should help speed-up string terms aggregations as well.The text was updated successfully, but these errors were encountered: