[GitHub] spark issue #19387: [SPARK-22160][SQL] Make sample points per partition (in ...
Github user xuanwang14 commented on the issue: https://github.com/apache/spark/pull/19387 LGTM --- - To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org
[GitHub] spark pull request #19387: [SPARK-22160][SQL] Make sample points per partiti...
Github user xuanwang14 commented on a diff in the pull request: https://github.com/apache/spark/pull/19387#discussion_r141763083 --- Diff: sql/core/src/test/scala/org/apache/spark/sql/ConfigBehaviorSuite.scala --- @@ -0,0 +1,64 @@ +/* + * 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.spark.sql + +import org.apache.commons.math3.stat.inference.ChiSquareTest + +import org.apache.spark.sql.internal.SQLConf +import org.apache.spark.sql.test.SharedSQLContext + + +class ConfigBehaviorSuite extends QueryTest with SharedSQLContext { + + import testImplicits._ + + test("SPARK-22160 spark.sql.execution.rangeExchange.sampleSizePerPartition") { +// In this test, we run a sort and compute the histogram for partition size post shuffle. +// With a high sample count, the partition size should be more evenly distributed, and has a +// low chi-sq test value. + +val numPartitions = 4 + +def computeChiSquareTest(): Double = { + val n = 1 + // Trigger a sort + val data = spark.range(0, n, 1, 1).sort('id) +.selectExpr("SPARK_PARTITION_ID() pid", "id").as[(Int, Long)].collect() + + // Compute histogram for the number of records per partition post sort + val dist = data.groupBy(_._1).map(_._2.length.toLong).toArray + assert(dist.length == 4) + + new ChiSquareTest().chiSquare( +Array.fill(numPartitions) { n.toDouble / numPartitions }, +dist) +} + +withSQLConf(SQLConf.SHUFFLE_PARTITIONS.key -> numPartitions.toString) { + // The default chi-sq value should be low + assert(computeChiSquareTest() < 100) + + withSQLConf(SQLConf.RANGE_EXCHANGE_SAMPLE_SIZE_PER_PARTITION.key -> "1") { +// If we only sample one point, the range boundaries will be pretty bad and the +// chi-sq value would be very high. +assert(computeChiSquareTest() > 1000) --- End diff -- This test may be flaky as well. --- - To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org
[GitHub] spark pull request #19387: [SPARK-22160][SQL] Make sample points per partiti...
Github user xuanwang14 commented on a diff in the pull request: https://github.com/apache/spark/pull/19387#discussion_r141762963 --- Diff: sql/core/src/test/scala/org/apache/spark/sql/ConfigBehaviorSuite.scala --- @@ -0,0 +1,64 @@ +/* + * 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.spark.sql + +import org.apache.commons.math3.stat.inference.ChiSquareTest + +import org.apache.spark.sql.internal.SQLConf +import org.apache.spark.sql.test.SharedSQLContext + + +class ConfigBehaviorSuite extends QueryTest with SharedSQLContext { + + import testImplicits._ + + test("SPARK-22160 spark.sql.execution.rangeExchange.sampleSizePerPartition") { +// In this test, we run a sort and compute the histogram for partition size post shuffle. +// With a high sample count, the partition size should be more evenly distributed, and has a +// low chi-sq test value. + +val numPartitions = 4 + +def computeChiSquareTest(): Double = { + val n = 1 + // Trigger a sort + val data = spark.range(0, n, 1, 1).sort('id) +.selectExpr("SPARK_PARTITION_ID() pid", "id").as[(Int, Long)].collect() + + // Compute histogram for the number of records per partition post sort + val dist = data.groupBy(_._1).map(_._2.length.toLong).toArray + assert(dist.length == 4) + + new ChiSquareTest().chiSquare( +Array.fill(numPartitions) { n.toDouble / numPartitions }, +dist) +} + +withSQLConf(SQLConf.SHUFFLE_PARTITIONS.key -> numPartitions.toString) { + // The default chi-sq value should be low + assert(computeChiSquareTest() < 100) --- End diff -- This test may be flaky. It depends on the ratio of `n/RANGE_EXCHANGE_SAMPLE_SIZE_PER_PARTITION`. What is the default value of `RANGE_EXCHANGE_SAMPLE_SIZE_PER_PARTITION` ? --- - To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org