While looking into mysql import error that I was facing (which was actually
my error), I came across following issue on jira -
This is about using avro file(s) instead of plain-text file(s) as
intermediate file while doing hive imports.
I can think about couple of scenarios where avro might help (but not very
sure) - binary data in database columns, data containing delimiters used
for field/line separators etc.
Has anyone tried implementing this? Any particular case that is solved by
using avro as intermediate format?