Optimize filters
execution another way
#88660
Labels
:Analytics/Aggregations
Aggregations
>enhancement
Team:Analytics
Meta label for analytical engine team (ESQL/Aggs/Geo)
Description
Back when I was working on #74260 I investigated another way of processing the
filters
agg by iterating the main query and thefilters
clauses in something similar feeling toDisjunctionScorer
that could remember which queries matched. I ultimately abandoned it because it was slower than the "native"terms
andrange
agg, but it was much faster than trying to execute either the filter by filter with very complex queries or the "compatible" mode for theterms
agg. It should be possible to bring that back, though I don't have the code anymore. If we find ourselves running frequently having to evaluatefilters
aggs at the top level that can not be otherwise rewritten into filter-by-filter, this could be a useful middle ground.The text was updated successfully, but these errors were encountered: