From 205e6d5867b180a85bad58035c917ca13552a0a5 Mon Sep 17 00:00:00 2001
From: Cheng Lian <lian@databricks.com>
Date: Wed, 9 Nov 2016 09:49:02 -0800
Subject: [PATCH] [SPARK-18338][SQL][TEST-MAVEN] Fix test case initialization
order under Maven builds
## What changes were proposed in this pull request?
Test case initialization order under Maven and SBT are different. Maven always creates instances of all test cases and then run them all together.
This fails `ObjectHashAggregateSuite` because the randomized test cases there register a temporary Hive function right before creating a test case, and can be cleared while initializing other successive test cases. In SBT, this is fine since the created test case is executed immediately after creating the temporary function.
To fix this issue, we should put initialization/destruction code into `beforeAll()` and `afterAll()`.
## How was this patch tested?
Existing tests.
Author: Cheng Lian <lian@databricks.com>
Closes #15802 from liancheng/fix-flaky-object-hash-agg-suite.
---
.../execution/ObjectHashAggregateSuite.scala | 23 ++++++++-----------
1 file changed, 10 insertions(+), 13 deletions(-)
diff --git a/sql/hive/src/test/scala/org/apache/spark/sql/hive/execution/ObjectHashAggregateSuite.scala b/sql/hive/src/test/scala/org/apache/spark/sql/hive/execution/ObjectHashAggregateSuite.scala
index 527626beeb..93fc5e8a5e 100644
--- a/sql/hive/src/test/scala/org/apache/spark/sql/hive/execution/ObjectHashAggregateSuite.scala
+++ b/sql/hive/src/test/scala/org/apache/spark/sql/hive/execution/ObjectHashAggregateSuite.scala
@@ -25,11 +25,10 @@ import org.scalatest.Matchers._
import org.apache.spark.sql._
import org.apache.spark.sql.catalyst.FunctionIdentifier
import org.apache.spark.sql.catalyst.analysis.UnresolvedFunction
-import org.apache.spark.sql.catalyst.expressions.{ExpressionEvalHelper, ExpressionInfo, Literal}
+import org.apache.spark.sql.catalyst.expressions.{ExpressionEvalHelper, Literal}
import org.apache.spark.sql.catalyst.expressions.aggregate.ApproximatePercentile
import org.apache.spark.sql.execution.aggregate.{HashAggregateExec, ObjectHashAggregateExec, SortAggregateExec}
import org.apache.spark.sql.functions._
-import org.apache.spark.sql.hive.HiveSessionCatalog
import org.apache.spark.sql.hive.test.TestHiveSingleton
import org.apache.spark.sql.internal.SQLConf
import org.apache.spark.sql.test.SQLTestUtils
@@ -43,6 +42,14 @@ class ObjectHashAggregateSuite
import testImplicits._
+ protected override def beforeAll(): Unit = {
+ sql(s"CREATE TEMPORARY FUNCTION hive_max AS '${classOf[GenericUDAFMax].getName}'")
+ }
+
+ protected override def afterAll(): Unit = {
+ sql(s"DROP TEMPORARY FUNCTION IF EXISTS hive_max")
+ }
+
test("typed_count without grouping keys") {
val df = Seq((1: Integer, 2), (null, 2), (3: Integer, 4)).toDF("a", "b")
@@ -199,10 +206,7 @@ class ObjectHashAggregateSuite
val typed = percentile_approx($"c0", 0.5)
// A Hive UDAF without partial aggregation support
- val withoutPartial = {
- registerHiveFunction("hive_max", classOf[GenericUDAFMax])
- function("hive_max", $"c1")
- }
+ val withoutPartial = function("hive_max", $"c1")
// A Spark SQL native aggregate function with partial aggregation support that can be executed
// by the Tungsten `HashAggregateExec`
@@ -420,13 +424,6 @@ class ObjectHashAggregateSuite
}
}
- private def registerHiveFunction(functionName: String, clazz: Class[_]): Unit = {
- val sessionCatalog = spark.sessionState.catalog.asInstanceOf[HiveSessionCatalog]
- val builder = sessionCatalog.makeFunctionBuilder(functionName, clazz.getName)
- val info = new ExpressionInfo(clazz.getName, functionName)
- sessionCatalog.createTempFunction(functionName, info, builder, ignoreIfExists = false)
- }
-
private def function(name: String, args: Column*): Column = {
Column(UnresolvedFunction(FunctionIdentifier(name), args.map(_.expr), isDistinct = false))
}
--
GitLab