Sqoop2 Avro Support

Question

The sqoop client argument support export as Avro format as,

--as-avrodatafile+

But when look in to Hue and Client Java API for Sqoop, it support only TEXT_FILE and SEQUENCE_FILE. how to enable support for AVRO and any documentation how to use it? Will Avro create schema based on the column names?

Answer 1

The difference here is that Sqoop have currently two major versions - Sqoop 1 and Sqoop 2. Sqoop 1 is an older stable version that do supports Avro format via the --as-avrodatafile argument. Sqoop 2 is a new version that has been completely redesigned from scratch to address new use cases and fix some security concerns. Sadly the Avro support is not there yet for Sqoop 2. As it happens Hue is using Sqoop 2 and hence you do not see option to import into Avro. You can still use Sqoop 1 if you need import into Avro though.