Troubleshooting ORC Tables

Issue

Spark pipeline does not display data for ORC source tables even when data exists in the table.

Cause

Vectorized reader is used when the spark.sql.hive.convertMetastoreOrc configuration in the /opt/infoworks/conf/df_spark_defaults.conf file is set to true.

Solution

This issue occurs in Spark 2.3 and higher versions.

Set the spark.sql.hive.convertMetastoreOrc configuration in the /opt/infoworks/conf/df_spark_defaults.conf file to false.