Fix OOM when data loading with the new Hive SNAPSHOT.

This patch is required for upgrading thirdparty.

After upgrading thirdparty, Hive OOMs in data loading TPCDS.
This patch fixes the OOM, but I have no idea why exactly it works.
I had tried several other approaches like increasing the
mapred and JVM heap memory to no avail.

Change-Id: I3c69e4c3bf0f24e0c3b6272c946f71c17e312e01
Reviewed-on: http://gerrit.cloudera.org:8080/2145
Reviewed-by: Alex Behm <alex.behm@cloudera.com>
Tested-by: Internal Jenkins

testdata/datasets/functional/functional_schema_template.sql

@@ -1694,7 +1694,7 @@ functional
 ---- BASE_TABLE_NAME
 avro_unicode_nulls
 ---- CREATE_HIVE
-create external table {db_name}{db_suffix}.{table_name} like {db_name}.liketbl stored as avro LOCATION '/test-warehouse/avro_null_char';
+create external table if not exists {db_name}{db_suffix}.{table_name} like {db_name}.liketbl stored as avro LOCATION '/test-warehouse/avro_null_char';
 ---- LOAD
 `hdfs dfs -mkdir -p /test-warehouse/avro_null_char && \
 hdfs dfs -put -f ${IMPALA_HOME}/testdata/avro_null_char/000000_0 /test-warehouse/avro_null_char/

testdata/datasets/tpcds/tpcds_schema_template.sql

@@ -463,8 +463,6 @@ WHERE 2452184 <= ss_sold_date_sk;
 ---- LOAD
 USE {db_name};
 
--- Disable auto.convert.join due to HIVE-5068
-set hive.auto.convert.join=false;
 set hive.exec.max.dynamic.partitions.pernode=10000;
 set hive.exec.max.dynamic.partitions=10000;
 set hive.exec.dynamic.partition.mode=nonstrict;
@@ -499,8 +497,6 @@ distribute by ss_sold_date_sk;
 ---- LOAD_LOCAL
 USE {db_name};
 
--- Disable auto.convert.join due to HIVE-5068
-set hive.auto.convert.join=false;
 set hive.exec.max.dynamic.partitions.pernode=10000;
 set hive.exec.max.dynamic.partitions=10000;
 set hive.exec.dynamic.partition.mode=nonstrict;