[SPARK-6775] Simplify CatalystConverter class hierarchy and pass in Parquet schema

• Type: Sub-task
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 1.0.2, 1.1.1, 1.2.1, 1.3.0
• Component/s: SQL
• Target Version/s: 1.5.0
• Assignee: Cheng Lian
• Reporter: Cheng Lian
• Created: 08/Apr/15 12:48
• Updated: 08/Jul/15 22:52
• Resolved: 08/Jul/15 22:52

Description

CataystConverter classes are used to convert Parquet records to Spark SQL row objects. Current converter implementations have the following problems:

1. They simply ignore original Parquet schema, which makes adding Parquet backwards-compatibility rules impossible.
2. They are unnecessary over complicated.
3. SpecificMutableRow is only used for structs whose fields are all of primitive types.

Issue Links

• links to: [Github] Pull Request #5422 (liancheng)
• links to: [Github] Pull Request #7231 (liancheng)