You'll have to define separate tables for the different schemas. You can "unify" them in a query with the union feature. You should also remove the header lines in the files, if you still have them, because Hive does not ignore them, but treats them as "data".
dean On Fri, Nov 30, 2012 at 2:59 AM, Marc Canaleta <mcanal...@gmail.com> wrote: > Hi all! > > We want to use hive to analyze our logs. Our logs will be TSV files, one > per hour, and as it is possible that we add/remove more columns in the > future, we will include headers (column names) in each file. > > So it is possible that two TSV files for different days/hours have > different headers. > > Is it possible to do this with Hive? > > Thanks! > -- *Dean Wampler, Ph.D.* thinkbiganalytics.com +1-312-339-1330
