File extensions not preserved in Hive.checkPaths when renaming new destination 
file
-----------------------------------------------------------------------------------

                 Key: HIVE-2407
                 URL: https://issues.apache.org/jira/browse/HIVE-2407
             Project: Hive
          Issue Type: Bug
    Affects Versions: 0.7.1
            Reporter: Brian Muller


In the checkPaths method of Hive.java, a new destination filename will be 
chosen if the source filename already exists in the destination directory.  
This new filename follows the simple schema of adding _copy_N to the source 
filename (with increasing N until a non-existing filename is found).

If the file has been LZO compressed it is imperative that the extension remain 
".lzo" so that the LZO indexer can find it and create a corresponding index 
file.  It would be much better to use a prefix of "copy_N_" or insert "_copy_N" 
somewhere else in the filename.

Without this, Hive and LZO compression will not work when identical source 
filenames are inserted into Hive.

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

Reply via email to