Yeah, more or less. I think that ticket is definitely a bigger issue, though in my case the subtle difference is that there simply are no columns AND no schema to go off of. The fix should probably be the number of columns when the schema is specified, and then go from there.
2011/12/19 Dmitriy Ryaboy <[email protected]> > Jon is this more or less https://issues.apache.org/jira/browse/PIG-2420 ? > > > On Mon, Dec 19, 2011 at 8:51 AM, Jonathan Coveney <[email protected]> > wrote: > > Imagine thing is just a file with 1\n2\n3\n4\n5 > > > > a = load 'thing' as (a:int); b = foreach (group a all) generate a, a as > b; > > c = foreach b generate FLATTEN(a) as a, FLATTEN({()}) ; dump c; > > > > This will return > > > > (1) > > (2) > > (3) > > (4) > > (5) > > > > The schema is unknown (perhaps an issue in and of itself). > > > > It FEELS like it should be: > > > > (1,) > > (2,) > > (3,) > > (4,) > > (5,) > > > > but I thought I'd get opinions. I have a patch that can do this, I'm just > > not sure if it's the desired outcome! >
