On Thu, Aug 17, 2017 at 11:18 AM, Piyush Mukati (Data Platform) < [email protected]> wrote:
> Hi, > Any thought on having config to overwrite the file with the same name > already exists (Line no 90 in org.apache.orc.impl.PhysicalFsWriter). > > https://issues.apache.org/jira/browse/ORC-231 > > Use Case: > I am using OrcOutputFormat with MultipleOutput in my MapReduce job. > Since MultipleOutput does not uses OutputCommitter so there are > half/corrupt files are left from the failed reducers. > > Here the files with the same name will be overwritten by the retry attempt > and it will guarantee correct result from a successful job. > > > Please suggest if any good production ready library available that can > replace MultipleOutputs and uses committer. > > thanks. >
ORC-231.patch
Description: Binary data
