[PIG-1330] Move pruned schema tracking logic from LoadFunc to core code

• Type: Improvement
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 0.7.0
• Fix Version/s: 0.7.0
• Component/s: impl
• Hadoop Flags: Reviewed
• Assignee: Daniel Dai
• Reporter: Daniel Dai
• Created: 25/Mar/10 22:13
• Updated: 14/May/10 06:47
• Resolved: 12/Apr/10 19:11

Description

Currently, LoadFunc.getSchema require a schema after column pruning. The good side of this is LoadFunc.getSchema matches the data it actually load. This gives a sense of consistency. However, by doing this, every LoadFunc need to keep track of the columns pruned. This is an unnecessary burden to the LoadFunc writer and it is very error proning. This issue is to move this logic from LoadFunc to Pig core. LoadFunc.getSchema then only need to return original schema even after pruning.

Issue Links

• relates to: PIG-1361 [Zebra] Zebra TableLoader.getSchema() should return the projectionSchema specified in the constructor of TableLoader instead of pruned proejction by pig