Hi,
I would like to ask a question on pig property to set the output directory
of MapReduce Job produced by pig. As we know,
"mapreduce.output.fileoutputformat.outputdir" or "mapred.output.dir" is the
property we use to set the output directory.
It seems that pig sets "pig.reduce.output.dirs" property in addition to the
property mentioned above.
They usually have the same value but I find that sometimes they don't (e.g
PigMix L12 - a case of multi stores), they are different. In this case,
hadoop follows "pig.reduce.output.dirs" property.
Could anyone explain how pig make hadoop use "pig.reduce.output.dirs"
property, instead of the usual property
("mapreduce.output.fileoutputformat.outputdir" or "mapred.output.dir"), to
set the output directory?
Please feel free to correct me if there is any false point in my question.
Thank you in advance.
Best Regards,
JI MAHN
