I think, PigServer is good to run. But, the script code place in java code, and I want to run dynamic pig script.
So, I use template engine like freemarker. Make pig script using template engine, and read it using StringReader. and modify the PigServer a little. 2014-02-17 6:10 GMT+09:00 Prashant Kommireddi <[email protected]>: > This might help > https://wiki.apache.org/pig/EmbeddedPig > > On Sunday, February 16, 2014, Jay Vyas <[email protected]> wrote: > > > Yeah that's what I think is the simplest: just exporting > > > > Hadoop_classpath=pig.jar:$hadoop_classpath > > > > Hadoop jar mymainclass arg1 arg2 > > > On Feb 16, 2014, at 3:04 PM, Prashant Kommireddi <[email protected] > <javascript:;>> > > wrote: > > > > > > You could run it as a regular java application (with a main method)? > You > > > need to have pig and Hadoop on the classpath. > > > > > >> On Sunday, February 16, 2014, Jay Vyas <[email protected] > <javascript:;>> > > wrote: > > >> > > >> Hi pig: > > >> > > >> What is the common idiom for executing a Java application which runs > pig > > >> commands using the direct Java API (i.e. by creatiing a PigServer > > object, > > >> etc...) ? There are a few ways i can think of: > > >> > > >> 1) using "hadoop jar" , but this will of course fail since pig is not > in > > >> the hadoop classpath. > > >> > > >> 2) using a "pig ..." command > > >> > > >> 3) Adding pig jars to the distributed cache at runtime in a standard > > java > > >> hadoop program. > > >> > > >> Not sure what is the "right" way to do it though? > > >> > > >> Thanks ! > > >> -- > > >> Jay Vyas > > >> http://jayunit100.blogspot.com > > >> > > >
