Switch to Lucene's new IntField/LongField/FloatField/DoubleField. #93165
Conversation
Lucene introduced new numeric fields that index both points and doc values. This has the same semantics as indexing one field for points and another one for doc values as we did before, but covering both data structures in a single field yielded a speedup in Lucene's nightly benchmarks (see annotation [AH](http://people.apache.org/~mikemccand/lucenebench/sparseResults.html#index_throughput)) which would be interesting to get too. This commit does not switch to factory methods for queries such as `LongField#newRangeQuery` for now, we'll need to look into it in a follow-up.
elasticsearchmachine
added
needs:triage
jpountz
added
>enhancement
:Search Foundations/Mapping
|
Hi @jpountz, I've created a changelog YAML for you. |
elasticsearchmachine
removed
the
needs:triage
|
Pinging @elastic/es-search (Team:Search) |
| } else { | ||
| context.addToFieldNames(fieldType().name()); | ||
| } |
There was a problem hiding this comment.
Previously, we only called addToFieldNames if at least store was true. Now, we call addToFieldNames no matter the store value. Is this intentional?
There was a problem hiding this comment.
Thanks for catching this, I had in my mind that we should add to field names whenever doc values is false, but this field and others do indeed only do this if the field is either indexed or stored as well.
|
the bwc test failure is one that has happened before, does not look like it's caused by your test, see #92058 . The percolator failure seems legit :) |
Absolutely, it's caused by apache/lucene#12109. The test failure should go away when we upgrade to the final release. |
|
@elasticmachine update branch |
|
@jpountz you were right about the percolator test failure, also auto-merge for the win here :) |
…astic#93165) Lucene introduced new numeric fields that index both points and doc values. This has the same semantics as indexing one field for points and another one for doc values as we did before, but covering both data structures in a single field yielded a speedup in Lucene's nightly benchmarks (see annotation [AH](http://people.apache.org/~mikemccand/lucenebench/sparseResults.html#index_throughput)) which would be interesting to get too. This commit does not switch to factory methods for queries such as `LongField#newRangeQuery` for now, we'll need to look into it in a follow-up.
Lucene introduced new numeric fields that index both points and doc values. This has the same semantics as indexing one field for points and another one for doc values as we did before, but covering both data structures in a single field yielded a speedup in Lucene's nightly benchmarks (see annotation AH) which would be interesting to get too.
This commit does not switch to factory methods for queries such as
LongField#newRangeQueryfor now, we'll need to look into it in a follow-up.