spark-instrumented-optimizer/sql/hive
Cheng Lian 81b35d8641 [SPARK-7684] [SQL] Invoking HiveContext.newTemporaryConfiguration() shouldn't create new metastore directory
The "Database does not exist" error reported in SPARK-7684 was caused by `HiveContext.newTemporaryConfiguration()`, which always creates a new temporary metastore directory and returns a metastore configuration pointing that directory. This makes `TestHive.reset()` always replaces old temporary metastore with an empty new one.

Author: Cheng Lian <lian@databricks.com>

Closes #6359 from liancheng/spark-7684 and squashes the following commits:

95d2eb8 [Cheng Lian] Addresses @marmbrust's comment
042769d [Cheng Lian] Don't create new temp directory in HiveContext.newTemporaryConfiguration()

(cherry picked from commit bfeedc69a2)
Signed-off-by: Cheng Lian <lian@databricks.com>
2015-05-26 00:16:24 +08:00
..
compatibility/src/test/scala/org/apache/spark/sql/hive/execution [SQL] [TEST] udf_java_method failed due to jdk version 2015-05-21 12:32:10 -07:00
src [SPARK-7684] [SQL] Invoking HiveContext.newTemporaryConfiguration() shouldn't create new metastore directory 2015-05-26 00:16:24 +08:00
v0.12.0/src/main/scala/org/apache/spark/sql/hive [SPARK-6638] [SQL] Improve performance of StringType in SQL 2015-04-15 13:06:38 -07:00
v0.13.1/src/main/scala/org/apache/spark/sql/hive [SPARK-6505] [SQL] Remove the reflection call in HiveFunctionWrapper 2015-04-27 14:08:05 +08:00
pom.xml Preparing development version 1.4.0-SNAPSHOT 2015-05-23 20:13:05 -07:00