0

I tried researching this but found no useful information. I have an M/R job already reading from parquet (not partitioned, using a thrift schema). I need to add another set of input files to the process that are not in parquet format, they're just regular csv files. Anyone know if this is possible or how it could be done?

Eric
  • 1
  • 1

1 Answers1

0

Never mind, I think i found what I needed in another post unrelated to parquet. Using multiple InputFormat classes while configuring MapReduce job

Here is the information I took from the answer I linked to and adapted to my own solution:

MultipleInputs.addInputPath(job, new Path("/path/to/parquet"), ParquetInputFormat.class, ParquetMapper.class);
MultipleInputs.addInputPath(job, new Path("/path/to/txt"), TextInputFormat.class, TextMapper.class);
Community
  • 1
  • 1
Eric
  • 1
  • 1
  • That's a classical [link-only answer](http://meta.stackoverflow.com/questions/251006/flagging-link-only-answers) which may be useless in a few weeks. Mind to extract the essence and improve your answer? – JensG Jan 25 '16 at 21:52