jacques-n commented on a change in pull request #2653: URL: https://github.com/apache/calcite/pull/2653#discussion_r774076758
########## File path: ubenchmark/src/jmh/java/org/apache/calcite/benchmarks/MetadataBenchmark.java ########## @@ -0,0 +1,140 @@ +/* + * 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.calcite.benchmarks; + +import org.apache.calcite.jdbc.Driver; +import org.apache.calcite.rel.RelNode; +import org.apache.calcite.rel.metadata.DefaultRelMetadataProvider; +import org.apache.calcite.rel.metadata.JaninoRelMetadataProvider; +import org.apache.calcite.rel.metadata.ProxyingMetadataHandlerProvider; +import org.apache.calcite.rel.metadata.RelMetadataQuery; +import org.apache.calcite.runtime.Hook; +import org.apache.calcite.test.CalciteAssert; + +import org.openjdk.jmh.annotations.Benchmark; +import org.openjdk.jmh.annotations.BenchmarkMode; +import org.openjdk.jmh.annotations.Fork; +import org.openjdk.jmh.annotations.Measurement; +import org.openjdk.jmh.annotations.Mode; +import org.openjdk.jmh.annotations.OutputTimeUnit; +import org.openjdk.jmh.annotations.Scope; +import org.openjdk.jmh.annotations.Setup; +import org.openjdk.jmh.annotations.State; +import org.openjdk.jmh.annotations.Threads; +import org.openjdk.jmh.annotations.Warmup; + +import java.sql.DriverManager; +import java.sql.SQLException; +import java.util.concurrent.TimeUnit; +import java.util.function.Consumer; +import java.util.function.Supplier; + +/** + * A benchmark to compare metadata retrieval time for a complex query. + * + * Compares metadata retrieval performance on a large query + * + */ +@Fork(value = 1, jvmArgsPrepend = "-Xmx2048m") +@State(Scope.Benchmark) +@Measurement(iterations = 10, time = 100, timeUnit = TimeUnit.MILLISECONDS) +@Warmup(iterations = 10, time = 100, timeUnit = TimeUnit.MILLISECONDS) +@Threads(1) +@OutputTimeUnit(TimeUnit.MILLISECONDS) +@BenchmarkMode(Mode.AverageTime) +public class MetadataBenchmark { + + @Setup + public void setup() throws SQLException { + DriverManager.registerDriver(new Driver()); + } + + private void test(final Supplier<RelMetadataQuery> supplier) { + CalciteAssert.that() + .with(CalciteAssert.Config.FOODMART_CLONE) + .query("select \"store\".\"store_country\" as \"c0\",\n" + + " \"time_by_day\".\"the_year\" as \"c1\",\n" + + " \"product_class\".\"product_family\" as \"c2\",\n" + + " count(\"sales_fact_1997\".\"product_id\") as \"m0\"\n" + + "from \"store\" as \"store\",\n" + + " \"sales_fact_1997\" as \"sales_fact_1997\",\n" + + " \"time_by_day\" as \"time_by_day\",\n" + + " \"product_class\" as \"product_class\",\n" + + " \"product\" as \"product\"\n" + + "where \"sales_fact_1997\".\"store_id\" = \"store\".\"store_id\"\n" + + "and \"store\".\"store_country\" = 'USA'\n" + + "and \"sales_fact_1997\".\"time_id\" = \"time_by_day\".\"time_id\"\n" + + "and \"time_by_day\".\"the_year\" = 1997\n" + + "and \"sales_fact_1997\".\"product_id\" = \"product\".\"product_id\"\n" + + "and \"product\".\"product_class_id\" = \"product_class\".\"product_class_id\"\n" + + "group by \"store\".\"store_country\",\n" + + " \"time_by_day\".\"the_year\",\n" + + " \"product_class\".\"product_family\"") + .withHook(Hook.CONVERTED, (Consumer<RelNode>) rel -> { + rel.getCluster().setMetadataQuerySupplier(supplier); + rel.getCluster().invalidateMetadataQuery(); + }) + .explainContains("" + + "EnumerableAggregate(group=[{1, 6, 10}], m0=[COUNT()])\n" + + " EnumerableMergeJoin(condition=[=($2, $8)], joinType=[inner])\n" + + " EnumerableSort(sort0=[$2], dir0=[ASC])\n" + + " EnumerableMergeJoin(condition=[=($3, $5)], joinType=[inner])\n" + + " EnumerableSort(sort0=[$3], dir0=[ASC])\n" + + " EnumerableHashJoin(condition=[=($0, $4)], joinType=[inner])\n" + + " EnumerableCalc(expr#0..23=[{inputs}], expr#24=['USA':VARCHAR(30)], " + + "expr#25=[=($t9, $t24)], store_id=[$t0], store_country=[$t9], $condition=[$t25])\n" + + " EnumerableTableScan(table=[[foodmart2, store]])\n" + + " EnumerableCalc(expr#0..7=[{inputs}], proj#0..1=[{exprs}], " + + "store_id=[$t4])\n" + + " EnumerableTableScan(table=[[foodmart2, sales_fact_1997]])\n" + + " EnumerableCalc(expr#0..9=[{inputs}], expr#10=[CAST($t4):INTEGER], " + + "expr#11=[1997], expr#12=[=($t10, $t11)], time_id=[$t0], the_year=[$t4], " + + "$condition=[$t12])\n" + + " EnumerableTableScan(table=[[foodmart2, time_by_day]])\n" + + " EnumerableHashJoin(condition=[=($0, $2)], joinType=[inner])\n" + + " EnumerableCalc(expr#0..14=[{inputs}], proj#0..1=[{exprs}])\n" + + " EnumerableTableScan(table=[[foodmart2, product]])\n" + + " EnumerableCalc(expr#0..4=[{inputs}], product_class_id=[$t0], " + + "product_family=[$t4])\n" + + " EnumerableTableScan(table=[[foodmart2, product_class]])") + .returns("c0=USA; c1=1997; c2=Non-Consumable; m0=16414\n" + + "c0=USA; c1=1997; c2=Drink; m0=7978\n" + + "c0=USA; c1=1997; c2=Food; m0=62445\n"); Review comment: I agree in theory. In reality, I'm looking for baby steps here. Based on my reviewing of metadata tickets back to the original refactor Julian did several years ago, it appears that multiple developers of the metadata module would manually use this query to test the performance. I wanted to at least get it in a place where it could be run in a reliable/standard/consistent way. Realistically, we should move to something that is actually a microbenchmark. Having the entirety of query planning rule firing/transformations/etc being part of a test to understand metadata performance leads to a lot of noise in the measurement (beyond just the plan validation). For now, I thought it was better to keep this in since it also serves to confirm that the providers are behaving consistently. -- 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: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
