Python UDF: import re causes instantiation error

[Devi Kumarappan <kp...@att.net>]

Hi,

I am using cduh3 distributionin the hadoop cluster . My python udf seems to work 
fine  if it does not import other python modules, but fails with ERROR 
org.apache.pig.tools.grunt.Grunt - ERROR 2997: Unable to recreate exception from 
backed error: java.io.IOException: Deserialization error: could not instantiate 
'org.apache.pig.scripting.jython.JythonFunction' with arguments  if I add 
"import re"  in the UDF. 

When I digged around i found that it has been fixed in pig0.10.
https://issues.apache.org/jira/browse/PIG-1824

But, CDUH3 has pig 0.8. Has anyody encountered this problem? Is there a work 
around  for this in Pig 0.8 or is it possible to update to pig 0.10 in CDUH3?

Thanks, Devi

Re: Python UDF: import re causes instantiation error

[Aniket Mokashi <an...@gmail.com>]

Put the re lib into the jython jar under Lib directory (or all of them), if
CDH has https://issues.apache.org/jira/browse/MAPREDUCE-967, then it should
work.

Also, to debug, import sys and print sys.path in your python script.

Thanks,
Aniket

On Mon, Apr 23, 2012 at 11:47 AM, Devi Kumarappan <kp...@att.net> wrote:

> Hi,
>
> I am using cduh3 distributionin the hadoop cluster . My python udf seems
> to work
> fine  if it does not import other python modules, but fails with ERROR
> org.apache.pig.tools.grunt.Grunt - ERROR 2997: Unable to recreate
> exception from
> backed error: java.io.IOException: Deserialization error: could not
> instantiate
> 'org.apache.pig.scripting.jython.JythonFunction' with arguments  if I add
> "import re"  in the UDF.
>
> When I digged around i found that it has been fixed in pig0.10.
> https://issues.apache.org/jira/browse/PIG-1824
>
> But, CDUH3 has pig 0.8. Has anyody encountered this problem? Is there a
> work
> around  for this in Pig 0.8 or is it possible to update to pig 0.10 in
> CDUH3?
>
> Thanks, Devi




-- 
"...:::Aniket:::... Quetzalco@tl"