Home | About | Sematext search-lucene.com search-hadoop.com
 Search Hadoop and all its subprojects:

Switch to Plain View
Pig, mail # user - Converting an inner bag


+
Josh Devins 2010-10-08, 16:13
Copy link to this message
-
Re: Converting an inner bag
Mehmet Tepedelenlioglu 2010-10-08, 17:12
I= foreach A generate group, flatten(items);

I believe that should do it.

On Oct 8, 2010, at 9:13 AM, Josh Devins wrote:

> I have a simple schema that contains an inner bag. What I need to
> essentially do is that for each tuple in the inner bag, I need to create a
> new tuple in a new outer bag. This is easier shown than explained! Consider
> the following schema and data:
>
> DESCRIBE A;
> A: {id: chararray, items: {item: chararray}}
>
> DUMP A;
> (a1, {(i1), (i2), (i3)})
> (a2, {(i1), (i2), (i3)})
>
> This needs to turn into some relation, I that looks like:
>
> DESCRIBE I;
> I: {id: chararray, item: chararray}
>
> DUMP I;
> (a1, i1)
> (a1, i2)
> (a1, i3)
> (a2, i1)
> (a2, i2)
> (a2, i3)
>
> ...so that I can then do a join with another relation X against the item
> field from I.
>
> Any ideas? Hope that description made sense.
>
> Many thanks,
>
> Josh
+
Josh Devins 2010-10-08, 21:00