hate to sound like a broken record but when all else fails think about the transform() function. The notion here is of encoding your tabs and newlines to something like '\t' and '\n' (literally) for instance. If those aren't unique enough use '<<tab>>' and "<<newline>>' (you get the idea) then having your app decode those strings to real tabs and real newlines when reading it.
What do you think? On Fri, May 3, 2013 at 2:07 AM, Valluri, Sathish <sathish.vall...@emc.com>wrote: > Hi All,**** > > ** ** > > We have an application which parses hive cli output and displays results.* > *** > > I have an external table with data in avro format, the contents in this > avro file have TAB and NEW LINES in the Avro data part.**** > > Since hive cli output rows are delimited by NEWLINES and columns are > delimited by TABS, if the actual content have TABS and NEW LINE characters > parsing the result set is giving wrong results.**** > > Can anyone suggest some ideas regarding delimiting the TABS and NEW LINE > characters in the hive cli output if the actual contents of the columns > have TABS and NEW LINES.**** > > ** ** > > Regards**** > > Sathish Valluri**** >
