hadoop-mapreduce-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Harsh J <ha...@cloudera.com>
Subject Re: output to file
Date Sun, 04 Dec 2011 17:14:37 GMT
Also, you don't need to run a reducer if you do not want one. Set your number of reduce tasks
to 0 before launching your job.

On 04-Dec-2011, at 9:03 PM, Praveen Sripati wrote:

> Matt,
> You can extend ArrayWritable. Also use TextOutputFormat as the output format.
> In the TextOutputFormat key.toString() and value.toString() are called, so override toString()
in the subclass of ArrayWritable to get the desired output format for the array. If toString()
is not overridden then class name and hex of the hash code is written to the output.
> Regards,
> Praveen
> On Sun, Dec 4, 2011 at 6:56 PM, Matt Denn <mattdenn12@yahoo.com> wrote:
> I have an array computed in map stage and want its ouptut recorded or sent to file. Reducer
can be naive and just send the output to file. Need to get a text output written to file for
this array.

View raw message