[PIG-2631] Pig should allow self joins

• Type: Improvement
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Fix Version/s: 0.18.0
• Assignee: Unassigned
• Reporter: Jonathan Coveney
• Created: 03/Apr/12 21:08
• Updated: 26/May/17 22:35

Description

This doesn't have to even be optimized, and can still involve a double scan of the data, but there is no reason the following should work:

a = load 'thing' as (x:int);
b = join a by x, (foreach a generate *) by x;

but this does not:

a = load 'thing' as (x:int);
b = join a by x, a by x;