com.intel.hadoop.graphbuilder.partition.mapreduce.vrecord
Class VrecordIngressMR
java.lang.Object
com.intel.hadoop.graphbuilder.partition.mapreduce.vrecord.VrecordIngressMR
public class VrecordIngressMR
- extends java.lang.Object
The MapRedue class takes from input directory a list of VertexRecrod
and distribute each record to its partitions. The output directory is
For example: this vrecord {gvid:"123","mirrors":[0,1,2,4],...} will appear in
the partition 0,1,2,4 of the output directory.
Input directory contains list of vertex records. Output directory structure:
- $outputdir/partition{$i}/vrecord for list of vertex records.
- $outputdir/partition{$i}/meta for meta info.
Method Summary |
void |
run(int numProcs,
java.lang.String inputpath,
java.lang.String outputpath)
|
void |
useGzip(boolean gzip)
Set compress option for the output, by default is false. |
Methods inherited from class java.lang.Object |
equals, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait |
VrecordIngressMR
public VrecordIngressMR()
useGzip
public void useGzip(boolean gzip)
- Set compress option for the output, by default is false.
- Parameters:
gzip
-
run
public void run(int numProcs,
java.lang.String inputpath,
java.lang.String outputpath)
throws java.io.IOException
- Throws:
java.io.IOException