You are viewing a plain text version of this content. The canonical link for it is here.
Posted to common-user@hadoop.apache.org by Грачев Роман <ro...@yandex.ru> on 2008/06/20 23:26:31 UTC

use recordio files as MapReduce input

Hi,

I used Hadoop recordio to write records into a file from the C++ program.
Now what I don't quite understand is how to use these files as an input to a MapReduce program.
Can someone please briefly describe the steps that I need to follow in order to do that?
Am I right in thinking that I need to implement a subclass of InputFormat, which produces a RecordReader, that is able to turn recordio files into key/value pairs?
Has anyone maybe done similar kind of thing already and can share a code example?

I am using recordio, as it provides a great way to read and write records both from C++ and Java, but I just don't see how it integrates into the rest of Hadoop.

Thanks,
Roman