Pig, mail # user - Results are different if I run a script or enter commands manually in grunt

Results are different if I run a script or enter commands manually in grunt

Uri Laserson 2013-01-25, 21:23

I am working on a Pig Script. One of the operations is computing a
quantile using Datafu's StreamingQuantile, and then using this value to
filter a relation. If I enter the commands one by one in the grunt shell,
the filtered relation contains some tuples. If I run the pig script from
bash, the filtered relation is empty. Where do I begin?

Thanks,
Uri

--
Uri Laserson, PhD
Data Scientist, Cloudera
Twitter/GitHub: @laserson
+1 617 910 0447
[EMAIL PROTECTED]

Re: Results are different if I run a script or enter commands manually in grunt

Cheolsoo Park 2013-01-25, 21:40

Hi Uri,

Can you turn off optimization when running a script and see if that returns
the same result as when you run commands one by one?
http://pig.apache.org/docs/r0.10.0/perf.html#optimization-rules

Thanks,
Cheolsoo

On Fri, Jan 25, 2013 at 1:23 PM, Uri Laserson <[EMAIL PROTECTED]> wrote:

> I am working on a Pig Script. One of the operations is computing a
> quantile using Datafu's StreamingQuantile, and then using this value to
> filter a relation. If I enter the commands one by one in the grunt shell,
> the filtered relation contains some tuples. If I run the pig script from
> bash, the filtered relation is empty. Where do I begin?
>
> Thanks,
> Uri
>
> --
> Uri Laserson, PhD
> Data Scientist, Cloudera
> Twitter/GitHub: @laserson
> +1 617 910 0447
> [EMAIL PROTECTED]
>