epotyom commented on code in PR #13568: URL: https://github.com/apache/lucene/pull/13568#discussion_r1693975540
########## lucene/sandbox/src/java/org/apache/lucene/sandbox/facet/ordinals/CandidateSetOrdinalIterator.java: ########## @@ -0,0 +1,48 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ +package org.apache.lucene.sandbox.facet.ordinals; + +import java.io.IOException; +import org.apache.lucene.internal.hppc.IntHashSet; + +/** + * {@link OrdinalIterator} that filters out ordinals from delegate if they are not in the candidate Review Comment: > What about this as an idea (maybe terrible... just brainstorming here)? What if we, (1) added a bool contains(int ord) method to FacetRecorder, and (2) changed this candidate iterator to take in a FacetRecorder during initialization instead of wrapping another iterator, then (3) it could "lead" the iteration from the candidate set provided and use contains to filter it down (rather than the current pattern that starts with all ordinals seen by the recorder and filters down using the provided set)? Good idea, it makes sense! Alternatively we can always require Count recorder, and check that the count for ordinal is greater than zero. But `contains` should be somewhat faster, and also I can't come up with a good reason why all FacetRecorders can't have `contains` method - so I'll implement `contains`. > This is of course assuming the common case is the recorder observes many more ordinals than provided candidates. Or maybe we could support both patterns somehow? Maybe the iterator could be smart about this like we are in other places ("leading" with the candidate or the recorder ordinals depending on heuristics)? Maybe we can also add `size` method to FacetRecorder interface? It can help make the decision - if `FacetRecorder.size` is greater than candidates size - use candidate ords as a source? It might be a premature optimization, in most cases candidate set should be smaller; so I suggest we implement the decision logic as a follow up. But we can implement `size` now, and remove `isEmpty` method, which is probably less useful? Parallel discussion about `isEmpty`: https://github.com/apache/lucene/pull/13568#discussion_r1691998941 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: issues-unsubscr...@lucene.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: issues-unsubscr...@lucene.apache.org For additional commands, e-mail: issues-h...@lucene.apache.org
