Embedding in Python in 0.9 is still done via Java object.  That is why we 
require Jython to execute the Python.  The major differences are that we 
presented a very simple interface (compile, bind, run) that was designed with 
iterative and parallel execution in mind.  We also made the bin/pig shell 
script that starts pig aware of Python so that the Python script itself can be 
passed to Pig.  This allows Pig to do things like return proper line numbers 
for error messages.

For a couple of great blogs on using the new interface, see 
http://www.hortonworks.com/pagerank-implementation-in-pig/ and 
http://www.hortonworks.com/transitive-closure-in-apache-pig/  For embedding 
docs see http://pig.apache.org/docs/r0.9.0/cont.html  For API docs on the new 
interface see 
http://pig.apache.org/docs/r0.9.0/api/org/apache/pig/scripting/Pig.html

Alan.

On Aug 24, 2011, at 8:07 AM, Shubham Chopra wrote:

> I am new to pig-0.9 and was exploring the embedding feature. I was wondering
> how is python embedding introduced in pig-0.9 different from embedded
> piglatin that could be used with java/scala? Any pointers to
> blogs/documentation about the same would be really great!
> 
> Thanks,
> Shubham.
> 
> On Fri, Jul 29, 2011 at 5:06 PM, Jeremy Hanna 
> <[email protected]>wrote:
> 
>> Nice work Daniel and all on the release and the blog posts!  Looking
>> forward to the other two.  We'll be testing out on our stuff because of all
>> the great features added.
>> 
>> On Jul 29, 2011, at 4:02 PM, Daniel Dai wrote:
>> 
>>> We wrote a serial of blogs to describe the new feature of Pig 0.9.0 on
>>> http://www.hortonworks.com/blog/. This serial contains three blogs and
>> will
>>> be published in a few days.
>>> 
>>> Thanks
>>> Daniel
>>> 
>>> On Fri, Jul 29, 2011 at 1:25 PM, Olga Natkovich <[email protected]>
>> wrote:
>>> 
>>>> Pig  team is happy to announce Pig 0.9.0 release.
>>>> 
>>>> Apache Pig provides a high-level data-flow language and execution
>> framework
>>>> for parallel computation on Hadoop clusters. More details about Pig can
>> be
>>>> found at http://pig.apache.org/.
>>>> 
>>>> The highlights of this release are introduction of control structures,
>>>> change of query parser, and semantic cleanup. The details of the release
>> can
>>>> be found at http://pig.apache.org/releases.html.
>>>> 
>>>> Olga
>>>> 
>>>> 
>> 
>> 

Reply via email to