[SPARK-23852] Parquet MR bug can lead to incorrect SQL results

• Type: Bug
• Status: Resolved
• Priority: Blocker
• Resolution: Fixed
• Affects Version/s: 2.3.0
• Fix Version/s: 2.3.1, 2.4.0
• Component/s: SQL
• Labels: correctness
• Assignee: Ryan Blue
• Reporter: Henry Robinson
• Created: 02/Apr/18 20:41
• Updated: 20/Sep/18 17:51
• Resolved: 10/May/18 02:56

Description

Parquet MR 1.9.0 and 1.8.2 both have a bug, PARQUET-1217, that means that pushing certain predicates to Parquet scanners can return fewer results than they should.

The bug triggers in Spark when:

• The Parquet file being scanner has stats for the null count, but not the max or min on the column with the predicate (Apache Impala writes files like this).
• The vectorized Parquet reader path is not taken, and the parquet-mr reader is used.
• A suitable <, <=, > or >= predicate is pushed down to Parquet.

The bug is that the parquet-mr interprets the max and min of a row-group's column as 0 in the absence of stats. So col > 0 will filter all results, even if some are > 0.

There is no upstream release of Parquet that contains the fix for PARQUET-1217, although a 1.10 release is planned.

The least impactful workaround is to set the Parquet configuration parquet.filter.stats.enabled to false.

Issue Links

• is caused by: PARQUET-1217 Incorrect handling of missing values in Statistics
• links to: [Github] Pull Request #21284 (henryr)
• links to: [Github] Pull Request #21302 (henryr)
• links to: [Github] Pull Request #21323 (henryr)