PigServer will connect to whichever cluster you point to in your configs. That is, embedded pig will launch job on the same cluster as would a regular Pig script.
Sent from my iPhone On Jan 27, 2012, at 9:58 AM, Michael <[email protected]> wrote: > When in embedded use, is it possible to create a PigServer instance which > connects to a remote hadoop namenode server? I'm assuming the PigServer > created in embedded mode executes job locally by default. > > > Thanks > > > > On Jan 28, 2012, at 1:40, Prashant Kommireddi <[email protected]> wrote: > >> Hi Vlad, >> >> Please take a look at >> http://wiki.apache.org/pig/EmbeddedPig >> >> Thanks, >> Prashant >> >> Sent from my iPhone >> >> On Jan 27, 2012, at 8:40 AM, Vlad S <[email protected]> wrote: >> >>> Hi Folks, >>> >>> While working with Pig it came to my mind that that currently there's >>> no way to call Pig Latin set of commands or a script from Java code >>> and receive results directly into Java. Or, maybe there's such way, >>> but I missed one - I'm newbie. >>> >>> I think having such functionality is as natural as having JDBC\ODBC >>> drivers to be able to execute SQL from the code and receive results in >>> the code. It might be beneficial for some systems that anticipate >>> script execution time to be reasonably short and script execution >>> outcome to be reasonably concise and predictable. >>> >>> Anyways, I implemented something relatively simple in a scope required >>> for my particular task(s). Let me know if such functionality is >>> desired - I'd be happy to contribute. Or, point me to some existing >>> functionality of such sort if I missed one. >>> >>> Regards, >>> Vlad.
