You are viewing a plain text version of this content. The canonical link for it is here.

Posted to user@hive.apache.org by Andrew O'Brien <ob...@gmail.com> on 2009/11/21 06:14:04 UTC

Partitioning from a single input

Hi everyone,

A question about partitioning: All of the examples I've seen insert
into a single hard-coded partition (usually a date) at a time.  If I
have a log file that spans a large range of dates, is there any way to
insert it into a partitioned table without having to explicitly
iterate through each date?  Something like PARTITION (log_date =
foo.date) (instead of PARTITION (log_date = '2009-11-21'))?

Thanks,
Andrew

Re: Partitioning from a single input

Posted by Chris Bates <ch...@gmail.com>.

A ticket has just been opened to add this functionality.  You should vote it
up!
https://issues.apache.org/jira/browse/HIVE-936

On Sat, Nov 21, 2009 at 12:14 AM, Andrew O'Brien <ob...@gmail.com>wrote:

> Hi everyone,
>
> A question about partitioning: All of the examples I've seen insert
> into a single hard-coded partition (usually a date) at a time.  If I
> have a log file that spans a large range of dates, is there any way to
> insert it into a partitioned table without having to explicitly
> iterate through each date?  Something like PARTITION (log_date =
> foo.date) (instead of PARTITION (log_date = '2009-11-21'))?
>
> Thanks,
> Andrew
>