[IMPALA-694] Parquet in CDH4.5 writes data files that are sometimes unreadable by Impala

• Type: Bug
• Status: Resolved
• Priority: Blocker
• Resolution: Fixed
• Affects Version/s: Impala 1.2
• Fix Version/s: Impala 1.2.2
• Assignee: Skye Wanderman-Milne
• Reporter: Nong Li
• Created: 10/Dec/13 21:06
• Updated: 12/Dec/13 23:43
• Resolved: 12/Dec/13 23:43

Description

The metadata written by parquet-mr is not correct which causes impala to be unable to read some files. This has been fixed in the latest version of parquet-mr (https://github.com/Parquet/parquet-mr/commit/aca1d8b895fb6f47e1e97e1f0133148ccbe4ab17) but we should consider making a compatibility fix in Impala to be able to handle those files.