java - Hadoop MultipleOutputs checksum -
I am using multiple outputs to write several files in Reducer. However, when I'm viewing in the target folder, I think the files have been successfully written, but the checksum file is empty (though made) is there any way to write checksum to multiple outputs?
I'm adding named output to the 'Run' method
Multiple output. AddNamedOutput (job, "fld1", TextOutputFormat class, text class, text class); ...
and write multiple outputs in reducers:
output = new multiple output (reference); ... output.write ("fld1", key, new text (arrays.toString (myData)));
Any ideas?
It appears that multiple optup object is required to be explicitly closed after typing. The following calls solve the problem: output.close ()
Comments
Post a Comment