From f22510cab592f13eea5e941e4e96cc19cf532192 Mon Sep 17 00:00:00 2001 From: Martijn van Groningen Date: Fri, 3 May 2013 18:04:35 +0200 Subject: [PATCH] A neater approach of for processing should clauses before must or must_not clauses. --- .../common/lucene/search/XBooleanFilter.java | 51 +++++++------------ 1 file changed, 18 insertions(+), 33 deletions(-) diff --git a/src/main/java/org/elasticsearch/common/lucene/search/XBooleanFilter.java b/src/main/java/org/elasticsearch/common/lucene/search/XBooleanFilter.java index d64999e4427..2221b794ec2 100644 --- a/src/main/java/org/elasticsearch/common/lucene/search/XBooleanFilter.java +++ b/src/main/java/org/elasticsearch/common/lucene/search/XBooleanFilter.java @@ -31,7 +31,9 @@ import org.elasticsearch.common.lucene.docset.DocIdSets; import org.elasticsearch.common.lucene.docset.NotDocIdSet; import java.io.IOException; -import java.util.*; +import java.util.ArrayList; +import java.util.Iterator; +import java.util.List; /** * Similar to {@link org.apache.lucene.queries.BooleanFilter}. @@ -78,7 +80,6 @@ public class XBooleanFilter extends Filter implements Iterable { boolean hasNonEmptyShouldClause = false; boolean hasMustClauses = false; boolean hasMustNotClauses = false; - boolean mustOrMustNotBeforeShould = false; for (int i = 0; i < clauses.size(); i++) { FilterClause clause = clauses.get(i); DocIdSet set = clause.getFilter().getDocIdSet(context, acceptDocs); @@ -88,9 +89,6 @@ public class XBooleanFilter extends Filter implements Iterable { return null; } } else if (clause.getOccur() == Occur.SHOULD) { - if (!hasShouldClauses && (hasMustClauses || hasMustNotClauses)) { - mustOrMustNotBeforeShould = true; - } hasShouldClauses = true; if (DocIdSets.isEmpty(set)) { continue; @@ -115,35 +113,11 @@ public class XBooleanFilter extends Filter implements Iterable { return null; } - if (mustOrMustNotBeforeShould) { - // Sort the clause only once if we encounter a should before a must or must_not clause - Collections.sort(clauses, new Comparator() { - @Override - public int compare(FilterClause o1, FilterClause o2) { - if (o1.getOccur() != o2.getOccur()) { - return o1.getOccur() == Occur.SHOULD ? -1 : 1; - } else { - return 0; - } - } - }); - - // Because we sorted the clause we also need to sort the result clauses - Collections.sort(results, new Comparator() { - @Override - public int compare(ResultClause o1, ResultClause o2) { - if (o1.clause.getOccur() != o2.clause.getOccur()) { - return o1.clause.getOccur() == Occur.SHOULD ? -1 : 1; - } else { - return 0; - } - } - }); - } - - // now, go over the clauses and apply the "fast" ones... + // now, go over the clauses and apply the "fast" ones first... hasNonEmptyShouldClause = false; boolean hasBits = false; + // But first we need to handle the "fast" should clauses, otherwise a should clause can unset docs + // that don't match with a must or must_not clause. for (int i = 0; i < results.size(); i++) { ResultClause clause = results.get(i); // we apply bits in based ones (slow) in the second run @@ -161,7 +135,18 @@ public class XBooleanFilter extends Filter implements Iterable { res = new FixedBitSet(reader.maxDoc()); } res.or(it); - } else if (clause.clause.getOccur() == Occur.MUST) { + } + } + + // Now we safely handle the "fast" must and must_not clauses. + for (int i = 0; i < results.size(); i++) { + ResultClause clause = results.get(i); + // we apply bits in based ones (slow) in the second run + if (clause.bits != null) { + hasBits = true; + continue; + } + if (clause.clause.getOccur() == Occur.MUST) { DocIdSetIterator it = clause.docIdSet.iterator(); if (it == null) { return null;