Flatten Component Does Not Work On Hive Parquet Storage in ODI 12.2.1.2.6 (Doc ID 2300840.1)

Last updated on SEPTEMBER 14, 2022

Applies to:

Symptoms

On Oracle Data Integrator (ODI) 12.2.1.2.6 when running a mapping where source is a Hive Parquet using "LKM Hive to Oracle OLH-OSCH Direct" and the Flatten component, the Flatten component is not able to explode the Hive Parquet and inject in the table. Even though the mapping execution completes successfully, no rows are inserted on target tables and the following messages can be observed in the ODI session logs:

17/06/30 16:05:43 INFO loader.OraLoader: Oracle Loader for Hadoop Release 3.8.0 - Production
Copyright (c) 2011, 2016, Oracle and/or its affiliates. All rights reserved.
17/06/30 16:05:43 INFO loader.OraLoader: Built-Against: hadoop-2.2.0hive-0.13.0 avro-1.8.1 jackson-1.8.8
17/06/30 16:05:52 INFO loader.OraLoader: oracle.hadoop.loader.loadByPartition is disabled because table: <TABLE NAME> is not partitioned
17/06/30 16:05:52 INFO loader.OraLoader: oracle.hadoop.loader.enableSorting disabled, no sorting key provided
17/06/30 16:05:52 INFO loader.OraLoader: Reduce tasks set to 0 because of no partitioning or sorting. Loading will be done in the map phase.
17/06/30 16:05:52 INFO output.DBOutputFormat: Setting map tasks speculative execution to false for : oracle.hadoop.loader.lib.output.JDBCOutputFormat

Here is a simple representation for an incorrect mapping which would produce the issue:

Changes

Cause

My Oracle Support provides customers with access to over a million knowledge articles and a vibrant support community of peers and Oracle experts.