Hi all,

I'm working on configuring a Hive deployment and looking for some guidance
on configuration. In hive-default.xml, there's a hive.default.fileformat
that currently defaults to text, but can also indicate sequence files. When
dealing with large volumes of data, shouldn't this default to sequence files
in general, and have users specifically indicate text for their final
outputs?

I'm told there might have been older issues around thrift data and
SequenceFiles, and that this might be fixed  now?

What do you all use for this setting?
Thanks,
- Aaron

Reply via email to