Yes, your thought was right! Using SequenceFileInputFormat should work fine (MapFile is just a specialization of it, a sorted one), so just pass the input paths to it.
On Tue, Jun 15, 2010 at 10:43 PM, Asif Jan <[email protected]> wrote: > Hi > > any pointers on how to use the MapFile with new mapreduce API. > > I did find the correspondinf output format e.g. > org.apache.hadoop.mapreduce.lib.output.MapFileOutputFormat, but was not able > to see how I can specify MapFileInputFormat ? (naively I thought that > org.apache.hadoop.mapreduce.lib.input.SequenceFileInputFormat; > should work for MapFile as well) > > will I have to implement RecordReader in order to read from a MapFile ? > > Thanks > > > > > > > > > > -- Harsh J www.harshj.com
