IMPALA-10871 (part 2): Apache Hive 3: fixes for dataset loading

This patch fixes the data loading problem of integrating Apache Hive 3
and switches to the tez engine.

Add HIVE-21569, HIVE-20038 patches and recompile the hive-exec module.

Testing:
- Manually perform data loading steps.

Change-Id: I86a1fdffc70b8d9a3bc97a72b5b939021dc496f1
Reviewed-on: http://gerrit.cloudera.org:8080/18028
Reviewed-by: Impala Public Jenkins <impala-public-jenkins@cloudera.com>
Tested-by: Impala Public Jenkins <impala-public-jenkins@cloudera.com>
This commit is contained in:
Fucun Chu
2021-11-15 19:31:30 +08:00
committed by Impala Public Jenkins
parent 06e8e7bba7
commit 1d16367afb
11 changed files with 381 additions and 2 deletions

View File

@@ -90,7 +90,8 @@ def load():
# explicitly creating an external table so that files are in the external warehouse
# directory. Use external.table.purge=true so that it is equivalent to a Hive 2
# managed table.
if HIVE_MAJOR_VERSION >= 3:
# For Apache Hive, HIVE-20085 (Hive 4) Allow CTAS.
if HIVE_MAJOR_VERSION >= 3 and os.environ["USE_APACHE_HIVE"] != "true":
external = "EXTERNAL"
tblproperties += ",'external.table.purge'='TRUE'"
sql_params = {