Sqoop

Apache Sqoop™ is a tool designed for efficiently transferring bulk data between Apache Hadoop and structured datastores such as relational databases.

This documentation provides all relevant details about using Sqoop and Sqoop2 with MapR, but does not duplicate Apache documentation. You can refer also to documentation available from the Apache Sqoop website.

The following table describes the differences between Sqoop1 or Sqoop2:

Feature Sqoop1 Sqoop2
Specialized connectors for all major RDBMS Available.

Not available. However, you can use the generic-jdbc-connector , which has been tested on these databases:

  • MySQL
  • Microsoft SQL Server
  • Oracle (Not supported in Sqoop 1.99.7)
  • PostgreSQL

The generic JDBC connector should also work with any other JDBC-compliant database, although specialized connectors probably give better performance.

Data transfer from RDBMS to Hive Done automatically.

Must be done manually in two stages:

  1. Import data from RDBMS into MapR-FS.
  2. Load data into Hive using the LOAD DATA command
    Note: As of Sqoop 1.99.7, you can also use the kite-connector to load data into Hive.
Data transfer from Hive to RDBMS

Must be done manually in two stages:

  1. Extract data from Hive into MapR-FS, as a text file or as an Avro file.
  2. Export the output of step 1 to an RDBMS using Sqoop.

Must be done manually in two stages:

  1. Extract data from Hive into MapR-FS, as a text file or as an Avro file.
    Note: As of Sqoop 1.99.7, you can also use the kite-connector to extract data from Hive.
  2. Export the output of step 1 to an RDBMS using Sqoop.
Integrated Kerberos security Supported. Supported.
Password encryption Not supported. Supported as of Sqoop 1.99.7.