gerlowskija commented on a change in pull request #1171: SOLR-13892: Add 'top-level' docValues Join implementation URL: https://github.com/apache/lucene-solr/pull/1171#discussion_r370910102
########## File path: solr/core/src/test/org/apache/solr/search/TestJoinQueryPerformance.java ########## @@ -0,0 +1,235 @@ +/* + * 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.solr.search; + +import java.io.BufferedReader; +import java.io.FileReader; +import java.util.Collection; +import java.util.HashMap; +import java.util.List; +import java.util.UUID; + +import com.google.common.collect.Lists; +import org.apache.solr.SolrTestCaseJ4; +import org.apache.solr.client.solrj.SolrQuery; +import org.apache.solr.client.solrj.SolrRequest; +import org.apache.solr.client.solrj.impl.HttpSolrClient; +import org.apache.solr.client.solrj.response.QueryResponse; +import org.apache.solr.common.SolrInputDocument; +import org.junit.BeforeClass; +import org.junit.Test; + + +public class TestJoinQueryPerformance extends SolrTestCaseJ4 { + // Dictionary used to load String data + private static final String DICT_PATH = "/usr/share/dict/words"; + private static final int NUM_DICT_TERMS = 235886; + private static final String[] LOADED_DICTIONARY = new String[NUM_DICT_TERMS]; + + // Performance run parameters: Indexing + private static final String FROM_COLLECTION_NAME = "user_acls"; + private static final int NUM_FROM_DOCS = 5050; // 1 + 2 + 3 + 4 + ... + 100 + private static final String TO_COLLECTION_NAME = "products"; + private static final int NUM_TO_DOCS = 500000; + private static final int PERMISSION_CARDINALITY = 50000; // 50K unique groups/roles/whatever + private static int BATCH_SIZE = 500; + private static int NUM_COMMITS = 500; + private static final int VAL_MAX = 1000; + private static final int USER_MAX = 100; + + private static String COLLECTION_NAME= "foo"; + + /* + * As I start out here, I think I'll want a few different axes. + * - "from" collection matches (with "to" matches held constant) + * - "to" collection matches (with "from" matches held constant) + * + * So I think I should index a finite number of docs + */ + + @BeforeClass + public static void setUpCluster() throws Exception { + loadDictionary(); + //loadCollectionData(DType.USER); + //loadCollectionData(DType.DATA); + } + + private static void loadDictionary() throws Exception { + try (BufferedReader reader = new BufferedReader(new FileReader(DICT_PATH))) { + for (int i = 0; i < NUM_DICT_TERMS; i++) { + LOADED_DICTIONARY[i] = reader.readLine(); + } + } + } + + public enum DType { + USER(NUM_FROM_DOCS, FROM_COLLECTION_NAME) { + // id - unique string + // userid_s - username (user# from 1-100)...each user appears in # entries + // permissions_ss - set of 300 string permissions (cardinality 50K) + @Override + SolrInputDocument buildDoc() { + if (userRecordCounts[currentUser - 1] == currentUser) { + currentUser++; + } else { + userRecordCounts[currentUser -1]++; + } + + final SolrInputDocument newDoc = new SolrInputDocument("id", UUID.randomUUID().toString()); + final String userString = "user" + currentUser; + final String[] permissions = getAFewDictionaryWords(300, PERMISSION_CARDINALITY); + + newDoc.addField("userid_s", userString); + newDoc.addField("permissions_ss", permissions); + + return newDoc; + } + }, + DATA(NUM_TO_DOCS, TO_COLLECTION_NAME) { + // id - unique string + // val_i - random int between 1-1000 + // cost_d - random cost between 1-1000 + // body_txt - random text string between 100 - 10000 words + // acl_ss - set of 100-3000 string permissions (cardinality 50K) + @Override + SolrInputDocument buildDoc() { + final SolrInputDocument newDoc = new SolrInputDocument("id", UUID.randomUUID().toString()); + final int val = random().nextInt(1000) + 1; + final double cost = random().nextDouble() * 1000d; + final String body = String.join(" ", getAFewDictionaryWords(random().nextInt(9900) + 100)); + final String[] acls = getAFewDictionaryWords(random().nextInt(2900) + 100, PERMISSION_CARDINALITY); + + newDoc.addField("val_i", val); + newDoc.addField("cost_d", cost); + newDoc.addField("body_txt", body); + newDoc.addField("acl_ss", acls); + + return newDoc; + } + }; + + private int numDocs; + private String collName; + private static int[] userRecordCounts = new int[100]; + private static int currentUser = 1; + + private DType(int numDocs, String collectionName) { + this.numDocs = numDocs; + this.collName = collectionName; + } + + abstract SolrInputDocument buildDoc(); + } + + private static void loadCollectionData(DType type) throws Exception { + int numDocs = type.numDocs; + String collectionName = type.collName; + int numLoaded = 0; + try (HttpSolrClient client = new HttpSolrClient.Builder("http://localhost:8983/solr";).build()) { + final int numBatches = numDocs / BATCH_SIZE + 1; + final int commitEveryBatches = NUM_COMMITS > 0 ? numBatches / NUM_COMMITS : Integer.MAX_VALUE; + int batchCount = 0; + while (numLoaded < numDocs) { + final int sizeOfBatch = numLoaded + BATCH_SIZE > numDocs ? numDocs - numLoaded : BATCH_SIZE; + final Collection<SolrInputDocument> batch = buildBatch(type, sizeOfBatch); + client.add(collectionName, batch); + batchCount++; + numLoaded+=sizeOfBatch; + + if (batchCount == commitEveryBatches) { + client.commit(collectionName); + batchCount = 0; + } + } + client.commit(collectionName); + } + + } + + private static Collection<SolrInputDocument> buildBatch(DType type, int sizeOfBatch) { + final List<SolrInputDocument> batch = Lists.newArrayList(); + for (int i = 0; i < sizeOfBatch; i++) { + batch.add(type.buildDoc()); + } + return batch; + } + + private static String[] getAFewDictionaryWords(int numWords) { + return getAFewDictionaryWords(numWords, NUM_DICT_TERMS); + } + + private static String[] getAFewDictionaryWords(int numWords, int onlyFirstN) { + final String[] words = new String[numWords]; + for (int i = 0; i < numWords; i++) { + words[i] = LOADED_DICTIONARY[random().nextInt(onlyFirstN)]; + } + + return words; + } + + + @Test + public void testJoinPerformanceAsMainQueryHitsIncrease() throws Exception { + final String joinQueryBase = "{!join fromIndex=" + FROM_COLLECTION_NAME + " from=permissions_ss to=acl_ss cache=false"; + final String fromQuery = "userid_s:user25"; // The higher the user number, the more permissions he has attached to his name (1-100) + final String standardJoin = joinQueryBase + "}" + fromQuery; + final String noScoreJoin = joinQueryBase + " score=none}" + fromQuery; + final String tpiJoin = joinQueryBase + " toplevel=true}" + fromQuery; + + try (HttpSolrClient client = new HttpSolrClient.Builder("http://localhost:8983/solr";).build()) { + for ( int i = 0; i < VAL_MAX; i+=20) { + final String mainQuery = "val_i:[1 TO " + (i+1) + "]"; + final QueryResponse standardJoinRsp = client.query(TO_COLLECTION_NAME, new SolrQuery("q", mainQuery, "fq", standardJoin), SolrRequest.METHOD.POST); + final QueryResponse tpiJoinRsp = client.query(TO_COLLECTION_NAME, new SolrQuery("q", mainQuery, "fq", tpiJoin), SolrRequest.METHOD.POST); + final QueryResponse noScoreJoinRsp = client.query(TO_COLLECTION_NAME, new SolrQuery("q", mainQuery, "fq", noScoreJoin), SolrRequest.METHOD.POST); + final long numFound = tpiJoinRsp.getResults().getNumFound(); + + System.out.println(i + "," + numFound + "," + standardJoinRsp.getQTime() + "," + noScoreJoinRsp.getQTime() + "," + tpiJoinRsp.getQTime()); + } + } + } + + @Test + public void testJoinPerformanceAsFromQueryHitsIncrease() throws Exception { + final String mainQuery = "val_i:[1 TO 250]"; // Half the docs match the query (250K) + + final String joinQueryBase = "{!join fromIndex=" + FROM_COLLECTION_NAME + " from=permissions_ss to=acl_ss cache=false"; + + + try (HttpSolrClient client = new HttpSolrClient.Builder("http://localhost:8983/solr";).build()) { + for ( int i = 1; i <= USER_MAX; i++) { + final String fromQuery = "userid_s:user" + i; + final String standardJoin = joinQueryBase + "}" + fromQuery; + final String noScoreJoin = joinQueryBase + " score=none}" + fromQuery; + final String tpiJoin = joinQueryBase + " toplevel=true}" + fromQuery; + + final QueryResponse standardJoinRsp = client.query(TO_COLLECTION_NAME, new SolrQuery("q", mainQuery, "fq", standardJoin), SolrRequest.METHOD.POST); + final QueryResponse tpiJoinRsp = client.query(TO_COLLECTION_NAME, new SolrQuery("q", mainQuery, "fq", tpiJoin), SolrRequest.METHOD.POST); + final QueryResponse noScoreJoinRsp = client.query(TO_COLLECTION_NAME, new SolrQuery("q", mainQuery, "fq", noScoreJoin), SolrRequest.METHOD.POST); + final long numFound = tpiJoinRsp.getResults().getNumFound(); + + System.out.println(i + "," + numFound + "," + standardJoinRsp.getQTime() + "," + noScoreJoinRsp.getQTime() + "," + tpiJoinRsp.getQTime()); Review comment: Hey Mikhail, thanks for the review. I have no intentions of committing this file. I kept it in the PR up to this point so that people could reproduce/tweak the perf test I ran. But I'm going to remove it prior to commit. (In fact, I thought I already had removed it, but there was one last "cleanup" commit I'd neglected to push until right now. You should see this file gone from the PR now.) ---------------------------------------------------------------- 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. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services --------------------------------------------------------------------- To unsubscribe, e-mail: issues-unsubscr...@lucene.apache.org For additional commands, e-mail: issues-h...@lucene.apache.org
