Hi folks,

I am running a hive script that currently takes in the variables T, T_PREV, etc.  These are used for, amonst other things, adding paritions.  For example:

alter table my_table add if not exists partition (dt='${T}');
alter table my_table add if not exists partition (dt='${T_PREV}');
The format of T, T_PREV, and subsequent variables is 'yyyy-MM-dd-HH'.  I'd like to pass in only T and calculate the others in the .q script.  In MySQL, I'd do this by setting a local variable (hiveconf domain?) based on a calculation (e.g., convert to unix_timestamp, subtract an hour, convert back to string).  But I can't seem to find a way to do this dynamically in Hive.  Am I missing something?

Sincerely,
Matt VonKip

The support is not there for this. Variables is just string
replacement in the CLI. We do not have select without where ' select
1+2' and we have no way to capture a result set and further process
it. Most people just do this work outside of hive.
On Thu, Feb 23, 2012 at 1:59 PM, Matt Vonkip <[EMAIL PROTECTED]> wrote:
> Hi folks,
>
> I am running a hive script that currently takes in the variables T, T_PREV,
> etc.  These are used for, amonst other things, adding paritions.  For
> example:
>
> alter table my_table add if not exists partition (dt='${T}');
> alter table my_table add if not exists partition (dt='${T_PREV}');
>
> The format of T, T_PREV, and subsequent variables is 'yyyy-MM-dd-HH'.  I'd
> like to pass in only T and calculate the others in the .q script.  In MySQL,
> I'd do this by setting a local variable (hiveconf domain?) based on a
> calculation (e.g., convert to unix_timestamp, subtract an hour, convert back
> to string).  But I can't seem to find a way to do this dynamically in Hive.
> Am I missing something?
>
> Sincerely,
> Matt VonKip