Partitioning an external hbase table

[MailingList <sl...@gmail.com>]

Is it possible to define partitions for a external table backed by Hbase? If so what is the proper syntax? 

I already have an external table backed by base and I'm finding that for even simple SELECT queries the load isn't getting evenly distributed a across the map tasks.  Some tasks see as few as a few hundred map input records while others receive more than a million.

Re: Partitioning an external hbase table

["kulkarni.swarnim@gmail.com" <ku...@gmail.com>]

Unfortunately I don't think there is a clean way to achieve that (atleast
not one that I know of). Your option at this point is to run your queries
with a WHERE clause so that the predicate behind the scenes gets converted
to a range scan and restricts the amount of data that is being getting
scanned.


On Wed, May 15, 2013 at 8:22 PM, MailingList
<sl...@gmail.com>wrote:

>  Is it possible to define partitions for a external table backed by Hbase?
> If so what is the proper syntax?
>
> I already have an external table backed by base and I'm finding that for
> even simple SELECT queries the load isn't getting evenly distributed a
> across the map tasks.  Some tasks see as few as a few hundred map input
> records while others receive more than a million.
>
>


-- 
Swarnim