Concept | Connections to SQL databases#

Watch the video

Supported databases#

Dataiku provides varying levels of support for connections to SQL databases. These support levels include full support, limited support, and no support.

For the latest details on database support, be sure to check SQL databases in the reference documentation.

Dataiku supports SQL connections at different levels, including full support, partial support and no support.

For unsupported databases, connections from Dataiku are still possible by using the appropriate Java Database Connectivity (JDBC) drivers. In such cases, reading from the database often works, but it is rare that writing to the database would work without some further configuration.

In-database computation#

Apart from reading and writing datasets in SQL databases, Dataiku can leverage in-database computation when rendering charts, executing visual recipes, and executing SQL code recipes. Leveraging the SQL execution engine can speed up computation!

A screen capture showing the in-database SQl recipe engine.

Defining a connection#

Setting up a connection to an SQL database involves two main steps:

  1. Install the JDBC driver for your database into the Dataiku data directory by copying the driver’s JAR file (and any dependencies) to the DATA_DIR/lib/jdbc folder.

    Note

    Note that DATA_DIR refers to the data directory where Dataiku is installed.

  2. From the Administration section of the Dataiku homepage, a user with admin rights can create a new connection by going to the Connections tab, selecting the appropriate connection type (for the particular SQL database), and filling in parameter values such as the database name, host, port, user, and password.

    The admin can also define other settings like authentication mode or permissions for user groups on this connection. For example, the admin can grant the data team permission to read only or to read and write from this connection.

Importing datasets from the SQL connection#

Once permissions have been given, and the connection has been set up and tested, you can return to your project and import datasets from this SQL connection.

To do so, you just need to select the table that you want to import from the list of tables in the datable and create the dataset.

See also

For more information, see SQL databases in the reference documentation.