I found the hadoop18.jar which pig use contains configuration files, such as
hadoop-site.xml , hadoop-default.xml. These files will be archived to
pig.jar when run ant jar.
And when I use pig in embed java way, run the following code snippet:
PigServer pig = new PigServer(ExecType.MAPREDUCE);
it will default load these configuration files. So it is not easy for me to
make configuration when I want to use my own hadoop configuration files. So
I think it's better not put these configruation to hadoop18.jar. Let the
user provide their own configuration files.
What do you think? Or Does there exists some ways to make configuration ?