Create weights lazily in filter and filters aggregation #26983
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Previous to this change the weights for the filter and filters aggregation were created in the `Filter(s)AggregatorFactory` which meant that they were created regardless of whether the aggregator actually collects any documents. This meant that for filters that are expensive to initialise, requests would not be quick when the query of the request was (or effectively was) a `match_none` query. This change maintains a single Weight instance for each filter across parent buckets but passes a weight supplier to the aggregator instances which will create the weight on first call and then return that instance for subsequent calls.
jimczi
approved these changes
Oct 12, 2017
| Query filter = filterBuilder.toFilter(context.getQueryShardContext()); | ||
| weight = contextSearcher.createNormalizedWeight(filter, false); | ||
| } catch (IOException e) { | ||
| throw new AggregationInitializationException("Failed to initialse filter", e); |
jasontedor
added a commit
that referenced
this pull request
Oct 12, 2017
* master: (35 commits) Create weights lazily in filter and filters aggregation (#26983) Use a dedicated ThreadGroup in rest sniffer (#26897) Fire global checkpoint sync under system context Update by Query is modified to accept short `script` parameter. (#26841) Cat shards bytes (#26952) Add support for parsing inline script (#23824) (#26846) Change default value to true for transpositions parameter of fuzzy query (#26901) Adding unreleased 5.6.4 version number to Version.java Rename TCPTransportTests to TcpTransportTests (#26954) Fix NPE for /_cat/indices when no primary shard (#26953) [DOCS] Fixed indentation of the definition list. Fix formatting in channel close test Check for closed connection while opening Clarify systemd overrides [DOCS] Plugin Installation for Windows (#21671) Painless: add tests for cached boxing (#24163) Don't detect source's XContentType in DocumentParser.parseDocument() (#26880) Fix handling of paths containing parentheses Allow only a fixed-size receive predictor (#26165) Add Homebrew instructions to getting started ...
colings86
added a commit
that referenced
this pull request
Oct 13, 2017
Previous to this change the weights for the filter and filters aggregation were created in the `Filter(s)AggregatorFactory` which meant that they were created regardless of whether the aggregator actually collects any documents. This meant that for filters that are expensive to initialise, requests would not be quick when the query of the request was (or effectively was) a `match_none` query. This change maintains a single Weight instance for each filter across parent buckets but passes a weight supplier to the aggregator instances which will create the weight on first call and then return that instance for subsequent calls.
colings86
added a commit
that referenced
this pull request
Oct 13, 2017
Previous to this change the weights for the filter and filters aggregation were created in the `Filter(s)AggregatorFactory` which meant that they were created regardless of whether the aggregator actually collects any documents. This meant that for filters that are expensive to initialise, requests would not be quick when the query of the request was (or effectively was) a `match_none` query. This change maintains a single Weight instance for each filter across parent buckets but passes a weight supplier to the aggregator instances which will create the weight on first call and then return that instance for subsequent calls.
colings86
added a commit
that referenced
this pull request
Oct 13, 2017
Previous to this change the weights for the filter and filters aggregation were created in the `Filter(s)AggregatorFactory` which meant that they were created regardless of whether the aggregator actually collects any documents. This meant that for filters that are expensive to initialise, requests would not be quick when the query of the request was (or effectively was) a `match_none` query. This change maintains a single Weight instance for each filter across parent buckets but passes a weight supplier to the aggregator instances which will create the weight on first call and then return that instance for subsequent calls.
jasontedor
added a commit
to olcbean/elasticsearch
that referenced
this pull request
Oct 16, 2017
* master: (356 commits) Do not set SO_LINGER on server channels (elastic#26997) Fix inconsistencies in the rest api specs for *_script (elastic#26971) fix inconsistencies in the rest api specs for cat.snapshots (elastic#26996) Add docs on full_id parameter in cat nodes API [TEST] Add test that replicates versioned updates with random flushes Use internal searcher for all indexing related operations in the engine Reformat paragraph in template docs to 80 columns Clarify settings and template on create index Fix reference to TcpTransport in documentation Allow Uid#decodeId to decode from a byte array slice (elastic#26987) Fix a typo in the similarity docs (elastic#26970) Use separate searchers for "search visibility" vs "move indexing buffer to disk (elastic#26972) Create weights lazily in filter and filters aggregation (elastic#26983) Use a dedicated ThreadGroup in rest sniffer (elastic#26897) Fire global checkpoint sync under system context Update by Query is modified to accept short `script` parameter. (elastic#26841) Cat shards bytes (elastic#26952) Add support for parsing inline script (elastic#23824) (elastic#26846) Change default value to true for transpositions parameter of fuzzy query (elastic#26901) Adding unreleased 5.6.4 version number to Version.java ...
jasontedor
added a commit
that referenced
this pull request
Oct 16, 2017
* 6.x: Remove unnecessary exception for engine constructor Update docs about `script` parameter (#27010) Do not set SO_LINGER on server channels (#26997) Fix inconsistencies in the rest api specs for *_script (#26971) fix inconsistencies in the rest api specs for cat.snapshots (#26996) Add docs on full_id parameter in cat nodes API Add removal of types to the 6.0 breaking changes Create weights lazily in filter and filters aggregation (#26983) [TEST] Add test that replicates versioned updates with random flushes Use internal searcher for all indexing related operations in the engine Use separate searchers for "search visibility" vs "move indexing buffer to disk (#26972) Reformat paragraph in template docs to 80 columns Clarify settings and template on create index Fix reference to TcpTransport in documentation Allow Uid#decodeId to decode from a byte array slice (#26987) Fix a typo in the similarity docs (#26970)
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Previous to this change the weights for the filter and filters aggregation were created in the
Filter(s)AggregatorFactorywhich meant that they were created regardless of whether the aggregator actually collects any documents. This meant that for filters that are expensive to initialise, requests would not be quick when the query of the request was (or effectively was) amatch_nonequery.This change maintains a single Weight instance for each filter across parent buckets but passes a weight supplier to the aggregator instances which will create the weight on first call and then return that instance for subsequent calls.