What are the various table creation arguments available in Sqoop?

Sqoop supports direct table creations in Hive, HBase and HCatalog as of now, and below are the corresponding create table arguments respectively.

• –create-hive-table argument used with –hive-import 
• –hbase-create-table along with –column-family <family> optionally –hbase-row-key <col>
• –create-hcatalog-table

In destination HBase table how will the row keys be maintained uniquely?

HBase tables maintain row keys uniquely and it is the primary key of input RDBMS table if it is present, otherwise it will be split-by column from input import command.

Can we bulk load HBase tables with RDBMS tables via Sqoop?

use of –hbase-bulkload argument enable bulk loading into HBase table while importing from RDBMS table via Sqoop.

Does Sqoop import direct fast importing technique while importing into Hive/HBase/Hcatalog?

No, fast direct import is supported only for HDFS and that too from MySQL and PostGreSQL only.

Can we import RDBMS tables into HCatalog directly, If yes, what are the limitations in it?

Yes by using –hcatalog-database option with –hcatalog-table we can create Hcatalog tables directly but importing/exporting into/from HCatalog has limitations as of now and doesn’t below arguments.

• –direct
• –export-dir
• –target-dir
• –warehouse-dir
• –append
• –as-sequencefile
• –as-avrofile

Can we load RDBMS tables data into an Hive Partition directly ? If Yes how can we achieve it?

Sqoop can import data for Hive into a particular partition by specifying the –hive-partition-keyand –hive-partition-value arguments.