[SPARK-34863] Support nested column in Spark Parquet vectorized readers

• Type: Sub-task
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 3.2.0
• Fix Version/s: 3.3.0
• Component/s: SQL
• Assignee: Chao Sun
• Reporter: Cheng Su
• Created: 25/Mar/21 08:57
• Updated: 03/Aug/22 03:00
• Resolved: 02/Apr/22 02:10

Description

The task is to support nested column type in Spark Parquet vectorized reader. Currently Parquet vectorized reader does not support nested column type (struct, array and map). We implemented nested column vectorized reader for FB-ORC in our internal fork of Spark. We are seeing performance improvement compared to non-vectorized reader when reading nested columns. In addition, this can also help improve the non-nested column performance when reading non-nested and nested columns together in one query.

 

Parquet: https://github.com/apache/spark/blob/master/sql/core/src/main/scala/org/apache/spark/sql/execution/datasources/parquet/ParquetFileFormat.scala#L173

Issue Links

• is a child of: SPARK-35743 Improve Parquet vectorized reader
• relates to: SPARK-39951 Support columnar batches with nested fields in Parquet V2
• links to: [Github] Pull Request #33695 (sunchao)
• links to: [Github] Pull Request #34659 (sunchao)
• links to: [Github] Pull Request #36055 (sunchao)
• links to: [Github] Pull Request #36123 (sunchao)
• links to: [Github] Pull Request #36123 (sunchao)
• links to: [Github] Pull Request #36366 (sadikovi)
• links to: [Github] Pull Request #36366 (sadikovi)
• links to: [Github] Pull Request #37379 (Kimahriman)