[DRILL-4156] Parquet group converter cannot cope with ENUM original type values

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 1.3.0, 1.4.0
• Fix Version/s: 1.13.0
• Component/s: Storage - Parquet
• Labels: easyfix newbie patch
• Flags: Patch
• Assignee: Unassigned
• Reporter: Chris Jansen
• Created: 03/Dec/15 12:51
• Updated: 26/Apr/18 15:52
• Resolved: 26/Apr/18 15:52

Description

When importing parquet data written using a Thrift schema, Drill cannot cope with ENUMs. This was fixed elsewhere in DRILL-1775 but not in the DrillParquetGroupConverter class (see mailing list thread).

It appears that the ParquetToDrillTypeConverter class was updated to use a varbinary as it's default behaviour, but the DrillParquetGroupConverter class was not updated to do the same.

Caused by: java.lang.UnsupportedOperationException: Unsupported type ENUM
	at org.apache.drill.exec.store.parquet2.DrillParquetGroupConverter.getConverterForType(DrillParquetGroupConverter.java:249)
	at org.apache.drill.exec.store.parquet2.DrillParquetGroupConverter.<init>(DrillParquetGroupConverter.java:154)
	at org.apache.drill.exec.store.parquet2.DrillParquetGroupConverter.<init>(DrillParquetGroupConverter.java:147)
	at org.apache.drill.exec.store.parquet2.DrillParquetGroupConverter.<init>(DrillParquetGroupConverter.java:147)
	at org.apache.drill.exec.store.parquet2.DrillParquetGroupConverter.<init>(DrillParquetGroupConverter.java:147)
	at org.apache.drill.exec.store.parquet2.DrillParquetRecordMaterializer.<init>(DrillParquetRecordMaterializer.java:40)
	at org.apache.drill.exec.store.parquet2.DrillParquetReader.setup(DrillParquetReader.java:267)
	... 14 more

Issue Links

• is part of: DRILL-5971 Fix INT64, INT32 logical types in complex parquet reader