Github user ioana-delaney commented on a diff in the pull request:
https://github.com/apache/spark/pull/15363#discussion_r107019056
--- Diff:
sql/catalyst/src/test/scala/org/apache/spark/sql/catalyst/optimizer/StarJoinReorderSuite.scala
---
@@ -0,0 +1,580 @@
+/*
+ * 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.catalyst.optimizer
+
+import org.apache.spark.sql.catalyst.SimpleCatalystConf
+import org.apache.spark.sql.catalyst.dsl.expressions._
+import org.apache.spark.sql.catalyst.dsl.plans._
+import org.apache.spark.sql.catalyst.expressions.{Attribute, AttributeMap}
+import org.apache.spark.sql.catalyst.plans.{Inner, PlanTest}
+import org.apache.spark.sql.catalyst.plans.logical.{ColumnStat,
LocalRelation, LogicalPlan}
+import org.apache.spark.sql.catalyst.rules.RuleExecutor
+import
org.apache.spark.sql.catalyst.statsEstimation.{StatsEstimationTestBase,
StatsTestPlan}
+
+
+class StarJoinReorderSuite extends PlanTest with StatsEstimationTestBase {
+
+ override val conf = SimpleCatalystConf(
+ caseSensitiveAnalysis = true, starSchemaDetection = true)
+
+ object Optimize extends RuleExecutor[LogicalPlan] {
+ val batches =
+ Batch("Operator Optimizations", FixedPoint(100),
+ CombineFilters,
+ PushDownPredicate,
+ ReorderJoin(conf),
+ PushPredicateThroughJoin,
+ ColumnPruning,
+ CollapseProject) :: Nil
+ }
+
+ // Table setup using star schema relationships:
+ //
+ // d1 - f1 - d2
+ // |
+ // d3 - s3
+ //
+ // Table f1 is the fact table. Tables d1, d2, and d3 are the dimension
tables.
+ // Dimension d3 is further joined/normalized into table s3.
+ // Tables' cardinality: f1 > d3 > d1 > d2 > s3
+ private val columnInfo: AttributeMap[ColumnStat] = AttributeMap(Seq(
+ // F1
+ attr("f1_fk1") -> ColumnStat(distinctCount = 3, min = Some(1), max =
Some(3),
+ nullCount = 0, avgLen = 4, maxLen = 4),
+ attr("f1_fk2") -> ColumnStat(distinctCount = 3, min = Some(1), max =
Some(3),
+ nullCount = 0, avgLen = 4, maxLen = 4),
+ attr("f1_fk3") -> ColumnStat(distinctCount = 4, min = Some(1), max =
Some(4),
+ nullCount = 0, avgLen = 4, maxLen = 4),
+ attr("f1_c4") -> ColumnStat(distinctCount = 4, min = Some(1), max =
Some(4),
+ nullCount = 0, avgLen = 4, maxLen = 4),
+ // D1
+ attr("d1_pk1") -> ColumnStat(distinctCount = 4, min = Some(1), max =
Some(4),
+ nullCount = 0, avgLen = 4, maxLen = 4),
+ attr("d1_c2") -> ColumnStat(distinctCount = 3, min = Some(1), max =
Some(3),
+ nullCount = 0, avgLen = 4, maxLen = 4),
+ attr("d1_c3") -> ColumnStat(distinctCount = 4, min = Some(1), max =
Some(4),
+ nullCount = 0, avgLen = 4, maxLen = 4),
+ attr("d1_c4") -> ColumnStat(distinctCount = 2, min = Some(2), max =
Some(3),
+ nullCount = 0, avgLen = 4, maxLen = 4),
+ // D2
+ attr("d2_c2") -> ColumnStat(distinctCount = 3, min = Some(1), max =
Some(3),
+ nullCount = 1, avgLen = 4, maxLen = 4),
+ attr("d2_pk1") -> ColumnStat(distinctCount = 3, min = Some(1), max =
Some(3),
+ nullCount = 0, avgLen = 4, maxLen = 4),
+ attr("d2_c3") -> ColumnStat(distinctCount = 3, min = Some(1), max =
Some(3),
+ nullCount = 0, avgLen = 4, maxLen = 4),
+ attr("d2_c4") -> ColumnStat(distinctCount = 2, min = Some(3), max =
Some(4),
+ nullCount = 0, avgLen = 4, maxLen = 4),
+ // D3
+ attr("d3_fk1") -> ColumnStat(distinctCount = 3, min = Some(1), max =
Some(3),
+ nullCount = 0, avgLen = 4, maxLen = 4),
+ attr("d3_c2") -> ColumnStat(distinctCount = 3, min = Some(1), max =
Some(3),
+ nullCount = 0, avgLen = 4, maxLen = 4),
+ attr("d3_pk1") -> ColumnStat(distinctCount = 5, min = Some(1), max =
Some(5),
+ nullCount = 0, avgLen = 4, maxLen = 4),
+ attr("d3_c4") -> ColumnStat(distinctCount = 2, min = Some(2), max =
Some(3),
+ nullCount = 0, avgLen = 4, maxLen = 4),
+ // S3
+ attr("s3_pk1") -> ColumnStat(distinctCount = 2, min = Some(1), max =
Some(2),
+ nullCount = 0, avgLen = 4, maxLen = 4),
+ attr("s3_c2") -> ColumnStat(distinctCount = 1, min = Some(3), max =
Some(3),
+ nullCount = 0, avgLen = 4, maxLen = 4),
+ attr("s3_c3") -> ColumnStat(distinctCount = 1, min = Some(3), max =
Some(3),
+ nullCount = 0, avgLen = 4, maxLen = 4),
+ attr("s3_c4") -> ColumnStat(distinctCount = 2, min = Some(3), max =
Some(4),
+ nullCount = 0, avgLen = 4, maxLen = 4),
+ // F11
+ attr("f11_fk1") -> ColumnStat(distinctCount = 3, min = Some(1), max =
Some(3),
+ nullCount = 0, avgLen = 4, maxLen = 4),
+ attr("f11_fk2") -> ColumnStat(distinctCount = 3, min = Some(1), max =
Some(3),
+ nullCount = 0, avgLen = 4, maxLen = 4),
+ attr("f11_fk3") -> ColumnStat(distinctCount = 4, min = Some(1), max =
Some(4),
+ nullCount = 0, avgLen = 4, maxLen = 4),
+ attr("f11_c4") -> ColumnStat(distinctCount = 4, min = Some(1), max =
Some(4),
+ nullCount = 0, avgLen = 4, maxLen = 4)
+ ))
+
+ private val nameToAttr: Map[String, Attribute] = columnInfo.map(kv =>
kv._1.name -> kv._1)
+ private val nameToColInfo: Map[String, (Attribute, ColumnStat)] =
+ columnInfo.map(kv => kv._1.name -> kv)
+
+ private val f1 = StatsTestPlan(
+ outputList = Seq("f1_fk1", "f1_fk2", "f1_fk3",
"f1_c4").map(nameToAttr),
+ rowCount = 6,
+ size = Some(48),
+ attributeStats = AttributeMap(Seq("f1_fk1", "f1_fk2", "f1_fk3",
"f1_c4").map(nameToColInfo)))
+
+ private val d1 = StatsTestPlan(
+ outputList = Seq("d1_pk1", "d1_c2", "d1_c3", "d1_c4").map(nameToAttr),
+ rowCount = 4,
+ size = Some(32),
+ attributeStats = AttributeMap(Seq("d1_pk1", "d1_c2", "d1_c3",
"d1_c4").map(nameToColInfo)))
+
+ private val d2 = StatsTestPlan(
+ outputList = Seq("d2_c2", "d2_pk1", "d2_c3", "d2_c4").map(nameToAttr),
+ rowCount = 3,
+ size = Some(24),
+ attributeStats = AttributeMap(Seq("d2_c2", "d2_pk1", "d2_c3",
"d2_c4").map(nameToColInfo)))
+
+ private val d3 = StatsTestPlan(
+ outputList = Seq("d3_fk1", "d3_c2", "d3_pk1", "d3_c4").map(nameToAttr),
+ rowCount = 5,
+ size = Some(40),
+ attributeStats = AttributeMap(Seq("d3_fk1", "d3_c2", "d3_pk1",
"d3_c4").map(nameToColInfo)))
+
+ private val s3 = StatsTestPlan(
+ outputList = Seq("s3_pk1", "s3_c2", "s3_c3", "s3_c4").map(nameToAttr),
+ rowCount = 2,
+ size = Some(17),
+ attributeStats = AttributeMap(Seq("s3_pk1", "s3_c2", "s3_c3",
"s3_c4").map(nameToColInfo)))
+
+ private val d3_ns = LocalRelation('d3_fk1.int, 'd3_c2.int, 'd3_pk1.int,
'd3_c4.int)
+
+ private val f11 = StatsTestPlan(
+ outputList = Seq("f11_fk1", "f11_fk2", "f11_fk3",
"f11_c4").map(nameToAttr),
+ rowCount = 6,
+ size = Some(48),
+ attributeStats = AttributeMap(Seq("f11_fk1", "f11_fk2", "f11_fk3",
"f11_c4")
+ .map(nameToColInfo)))
+
+ private val subq = d3.select(sum('d3_fk1).as('col))
+
+ test("Test 1: Selective star-join on all dimensions") {
+ // Star join:
+ // (=) (=)
+ // d1 - f1 - d2
+ // | (=)
+ // s3 - d3
+ //
+ // Query:
+ // select f1_fk1, f1_fk3
+ // from d1, d2, f1, d3, s3
+ // where f1_fk2 = d2_pk1 and d2_c2 < 2
+ // and f1_fk1 = d1_pk1
+ // and f1_fk3 = d3_pk1
+ // and d3_fk1 = s3_pk1
+ //
+ // Positional join reordering: d1, f1, d2, d3, s3
+ // Star join reordering: f1, d2, d1, d3, s3
+ val query =
+ d1.join(d2).join(f1).join(d3).join(s3)
+ .where((nameToAttr("f1_fk2") === nameToAttr("d2_pk1")) &&
+ (nameToAttr("d2_c2") === 2) &&
+ (nameToAttr("f1_fk1") === nameToAttr("d1_pk1")) &&
+ (nameToAttr("f1_fk3") === nameToAttr("d3_pk1")) &&
+ (nameToAttr("d3_fk1") === nameToAttr("s3_pk1")))
+
+ val expected =
+ f1.join(d2.where(nameToAttr("d2_c2") === 2), Inner,
+ Some(nameToAttr("f1_fk2") === nameToAttr("d2_pk1")))
+ .join(d1, Inner, Some(nameToAttr("f1_fk1") ===
nameToAttr("d1_pk1")))
+ .join(d3, Inner, Some(nameToAttr("f1_fk3") ===
nameToAttr("d3_pk1")))
+ .join(s3, Inner, Some(nameToAttr("d3_fk1") ===
nameToAttr("s3_pk1")))
+
+ assertEqualPlans(query, expected)
+ }
+
+ test("Test 2: Star join on a subset of dimensions due to inequality
joins") {
+ // Star join:
+ // (=) (<)
+ // d1 - f1 - d2
+ // |
+ // | (=)
+ // d3 - s3
+ // (=)
+ //
+ // Query:
+ // select f1_fk1, f1_fk3
+ // from d1, f1, d2, s3, d3
+ // where f1_fk2 < d2_pk1
+ // and f1_fk1 = d1_pk1 and d1_c2 = 2
+ // and f1_fk3 = d3_pk1
+ // and d3_fk1 = s3_pk1
+ //
+ // Default join reordering: d1, f1, d2, d3, s3
+ // Star join reordering: f1, d1, d3, d2,, d3
+
+ val query =
+ d1.join(f1).join(d2).join(s3).join(d3)
+ .where((nameToAttr("f1_fk2") < nameToAttr("d2_pk1")) &&
+ (nameToAttr("f1_fk1") === nameToAttr("d1_pk1")) &&
+ (nameToAttr("d1_c2") === 2) &&
+ (nameToAttr("f1_fk3") === nameToAttr("d3_pk1")) &&
+ (nameToAttr("d3_fk1") === nameToAttr("s3_pk1")))
+
+ val expected =
+ f1.join(d1.where(nameToAttr("d1_c2") === 2), Inner,
+ Some(nameToAttr("f1_fk1") === nameToAttr("d1_pk1")))
+ .join(d3, Inner, Some(nameToAttr("f1_fk3") ===
nameToAttr("d3_pk1")))
+ .join(d2, Inner, Some(nameToAttr("f1_fk2") < nameToAttr("d2_pk1")))
+ .join(s3, Inner, Some(nameToAttr("d3_fk1") ===
nameToAttr("s3_pk1")))
+
+ assertEqualPlans(query, expected)
+ }
+
+ test("Test 3: Star join on a subset of dimensions since join column is
not unique") {
+ // Star join:
+ // (=) (=)
+ // d1 - f1 - d2
+ // | (=)
+ // d3 - s3
+ //
+ // Query:
+ // select f1_fk1, f1_fk3
+ // from d1, f1, d2, s3, d3
+ // where f1_fk2 = d2_c4
+ // and f1_fk1 = d1_pk1 and d1_c2 = 2
+ // and f1_fk3 = d3_pk1
+ // and d3_fk1 = s3_pk1
+ //
+ // Default join reordering: d1, f1, d2, d3, s3
+ // Star join reordering: f1, d1, d3, d2, d3
--- End diff --
@cloud-fan Yes, it's a typo like above. I did some small changes to the
queries when I rewrote the test suite and didn't update the code comments
properly. I will fix. Thanks!
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
