histo -i in1 -o out1.file out1.mr
histo -i mrdata -o NULL out1.mr
This is a named command which calculates the frequency of key occurrence in an input set of key/values.
See the named command doc page for various ways in which the -i inputs and -o outputs for a named command can be specified.
In1 is any MR-MPI object containing key/value pairs. The input is unchanged by this command.
In1 cannot be specified as input from file(s) since no assumption is made about how the files store key/value pairs. You would need to read the files into a MR-MPI object as a pre-processing step, using a map() function you provide, before passing that object to the histo command as an input.
Out1 will store the frequency count of all unique keys.
Out1 cannot be specified as output to a file since no assumption is made about how the key should be formatted for output. You would need to write the files from the output MR-MPI object as a post-processing step, using a map() or scan() function you provide.
Statistics on the count of each key will be printed to the screen in sorted order.