[SPARK-22120][SQL] TestHiveSparkSession.reset() should clean out Hive…

… warehouse directory ## What changes were proposed in this pull request? During TestHiveSparkSession.reset(), which is called after each TestHiveSingleton suite, we now delete and recreate the Hive warehouse directory. ## How was this patch tested? Ran full suite of tests locally, verified that they pass. Author: Greg Owen <[email protected]> Closes apache#19341 from GregOwen/SPARK-22120.
ianlcsd · Sep 25, 2017 · ce20478 · ce20478
1 parent 038b185
commit ce20478
Showing 1 changed file with 6 additions and 0 deletions.
diff --git a/sql/hive/src/main/scala/org/apache/spark/sql/hive/test/TestHive.scala b/sql/hive/src/main/scala/org/apache/spark/sql/hive/test/TestHive.scala
@@ -18,6 +18,7 @@
 package org.apache.spark.sql.hive.test
 
 import java.io.File
+import java.net.URI
 import java.util.{Set => JavaSet}
 
 import scala.collection.JavaConverters._
@@ -498,6 +499,11 @@ private[hive] class TestHiveSparkSession(
         }
       }
 
+      // Clean out the Hive warehouse between each suite
+      val warehouseDir = new File(new URI(sparkContext.conf.get("spark.sql.warehouse.dir")).getPath)
+      Utils.deleteRecursively(warehouseDir)
+      warehouseDir.mkdir()
+
       sharedState.cacheManager.clearCache()
       loadedTables.clear()
       sessionState.catalog.reset()