Renato, slides 20-23 on this presentation describe the implementations of
the joins:
http://squarecog.wordpress.com/2009/11/03/apache-pig-apittsburgh-hadoop-user-group/

You can also search the pig wiki for design docs for each join.

D

On Tue, Apr 12, 2011 at 9:48 PM, Renato Marroquín Mogrovejo <
[email protected]> wrote:

> Hi there,
>
> I have some questions about how PIG performs joins. The site says there are
> three types of specialized joins: Replicated, skew, and merge joins. I
> wanted to know these implementations. For instance, what about the regular
> type of join?  Is there a difference between the repartition join
> (org.apache.hadoop.contrib.utils.join) and the regular join implemented in
> PIG? Is the merge join what is called a map-side join (where input is
> sorted
> and just regular scans are done)?
> And about PIG join framework[1], are there other specific implementation
> done? e.g. Semi-joins?
> Thanks in advance.
>
>
> Renato M.
>

Reply via email to