Begin forwarded message:
From: Milind A Bhandarkar
Date: May 31, 2010 9:16:38 PM PDT
To: "common-u...@hadoop.apache.org" ,
"mapreduce-u...@hadoop.apache.org" u...@hadoop.apache.org>, "gene...@hadoop.apache.org" >
Subject: First International Mapreduce Workshop 2010: Paper
submission deadl
I've created https://issues.apache.org/jira/browse/PIG-1459 to capture
the need for a standard serialization method.
Regarding required field list, it is the last option. I believe the
name was included since some loaders may think in terms of names
instead of positions. I created https:/
yeah, that'd be really cool. The other way that we can say this, (to make
map reduce interface available in pig), is to allow FOREACH to be nested:
TRIMED_TABLE = FOREACH TABLE {
stripped = FOREACH TABLE.SOME_BAG GENERATE String.Trim(value);
GENERATE k1,k2,k3, stripped;
}
On Wed, Ju
Another way to make tuples from other tuples or fields: use special UDF
https://issues.apache.org/jira/browse/PIG-1385
It depends mostly on what you need.
2010/6/18 Corbin Hoenes
> Need to have a bunch of non related aliases into a single alias (so I can
> pass this alias into my UDF). Is it
Tuples can contain tuples.
The arguments to a udf get wrapped in a tuple. So, MyUDF('foo', 'bar', 24)
will invoke MyUDF.exec(Tuple input), and this input tuple will have 3 fields
-- String foo, String bar, and int 24.
If you call a udf on a list of tuples -- MyUDF(tuple1, tuple2) -- this will
res
Need to have a bunch of non related aliases into a single alias (so I can pass
this alias into my UDF). Is it possible to do this? Or is it possible to pass
a number of Tuple objects into an EvalFunc? exec(Tuple input1, Tuple input2,
...)
