You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@pig.apache.org by zjffdu <zj...@gmail.com> on 2009/04/27 17:31:54 UTC

Why there's no Outer Join ?

 

Since there's outer cogroup, so why there's no outer join ?

 

And I think outer join is a must-have. 

 

 

Thank you

Jeff Zhang

 


Re: Why there's no Outer Join ?

Posted by Alan Gates <ga...@yahoo-inc.com>.
Flatten gets rid of rows where one of the bags is empty, so it will  
turn an outer cogroup into an inner join.

As to why there's no outer join, two reasons.  One, it isn't always  
possible to do an outer join in pig because an outer join requires  
that you know the schema of each input (otherwise you don't know how  
many nulls to add to the record).  Given this, we could still  
implement it for cases where both input schemas are known.  We do have  
plans to do that but haven't gotten there yet.

Alan.

On Apr 27, 2009, at 8:46 AM, Dmitriy Ryaboy wrote:

> Is an outer join different from an outer cogroup followed by a  
> Flatten?
>
> On Mon, Apr 27, 2009 at 11:31 AM, zjffdu <zj...@gmail.com> wrote:
>>
>>
>> Since there's outer cogroup, so why there's no outer join ?
>>
>>
>>
>> And I think outer join is a must-have.
>>
>>
>>
>>
>>
>> Thank you
>>
>> Jeff Zhang
>>
>>
>>
>>


Re: Why there's no Outer Join ?

Posted by Dmitriy Ryaboy <dv...@gmail.com>.
Is an outer join different from an outer cogroup followed by a Flatten?

On Mon, Apr 27, 2009 at 11:31 AM, zjffdu <zj...@gmail.com> wrote:
>
>
> Since there's outer cogroup, so why there's no outer join ?
>
>
>
> And I think outer join is a must-have.
>
>
>
>
>
> Thank you
>
> Jeff Zhang
>
>
>
>