Hi,
I tried to ran a cascading job with 2 input files and used a hashjoin pipe.I have created custom counters using each pipe to get the( Record Count) number of tuples flowing through a pipe/input.1st input file connected on lhs generated correct record count whereas for 2nd input file on rhs was not giving any record count.
After debugging the code, The getReporter method in HadoopFlowProcess class is returning two different reporter objects for the 2 input files, TaskReporter for lhs and NULL reporter for rhs.But we are expecting TaskReporter for both input files.
For the input file connected on lhs of Hashjoin, TaskReporter is incrementing the custom counter by incrCounter method and
for input file connected on rhs, Reporter is called to increment the counter(But this class have empty incrCounter method).
The above problem was faced when using Hadoop2MR1FlowConnector but when I ran the same job using Hadoop2TezFlowConnector, the
counters were successfully generated for both the inputs.
Can someone please provide any solution for this problem?
Cascading version : 3.1.0
Hadoop version : 2.6
Thanks