Hello, I am writing a MR job where the distribution of the Keys emitted by the Map phase is not known beforehand and so I can't create the partitions for the TotalOrderPartitioner. I would like to sample those keys to create the partitions and then run the job that will process the whole input.
Is the InputSampler the tool I need? I tried to use it but I think it doesn't use the mapper class to process the samples and then create the partitions, but it just creates the partitions from the input. Am I wrong? Thank you in advance! Pan