Hadoop Random Notes

Do not write, the default is the following conversion class
Job.setinputformatclass (Textinputformat.class)

List<inputsplit> Inputformat.getsplits first to the input data to do the segmentation, the split after splitting the written decision map task number;

Recordreader<k,v> Inputformat.createrecordreader (Inputsplit split, ...) Incoming shard data, processed into key, value, and then send keyvalue value to map execution, each pair of key, value pair will call a map;

Fileinputformat<k, v> extends Inputformat<k, v>

List<inputsplit> fileinputformat.getsplits

Textinputformat extends Fileinputformat<longwritable, text>

Hadoop Random Notes