Data Connections

Explore resources for working with connections to data sources from a user’s perspective.

Tip

To validate your knowledge of this area, register for the Academy course Dataiku & SQL, an optional part of the Core Designer learning path.

Concepts & tutorials

• Concept | Data connections
• Concept | Architecture model for databases
• Tutorial | Data transfer with the Sync recipe
• Tutorial | Data transfer with the Prepare recipe
• Concept | Connection changes
• Tutorial | Integration with Amazon Redshift
• Tutorial | Integration with MongoDB

Reference | A primer on connecting to data sources

Connect to your existing infrastructure

SQL databases

The list of supported SQL databases, and information on how to connect to them, is available from our documentation on SQL datasets.

Hadoop HDFS

To connect you will first need to configure Hadoop on your instance. Please refer to the detailed guides on specific Hadoop distributions and managed services.

NoSQL

• Cassandra

• MongoDB

• ElasticSearch

Accessing cloud storage and databases

Cloud File Storage

• Amazon S3

• Google Cloud Storage

• Azure Blob Storage

Cloud Databases

• Amazon Redshift. Syncing from S3 to Redshift is most efficient, and Dataiku DSS takes this route whenever possible, see this page for details.

• Google BigQuery is available through a JDBC driver developed by Simba.

• Snowflake

Fetching data from remote sources

It is possible to fetch data using various protocols, and caching the resulting dataset on the filesystem.

• You can use an FTP server as an uncached database.

• You can also fetch data using HTTP, and use secured connections through SFTP or SCP.

File formats

Dataiku can read and write in various file formats for files-based connections: filesystem, HDFS, Amazon S3, HTTP, FTP, SSH… See the list of readable file formats.

Accessing data through plugins

Many applications such as Google Sheets, SalesForce, Slack… provide capabilities to access their data through APIs. Dataiku plugins allow the addition of custom connections leveraging these APIs to easily define datasets that fetch data from a wide variety of applications.

See the available plugins or create your own plugin.

How-to | Utilize MS Access

Many of our users have shown interest in utilizing MS Access in Dataiku. In the interest of knowledge sharing, we wanted to demonstrate how to do just that.

How to open an MS Access file

• Download ucanaccess

• Copy ucanaccess-4.0.4.jar and jackcess-2.1.11.jar (including the ones from lib/) into DATA_DIR/lib/jdbc/ (see Installing database drivers for more details)

• Restart DSS.

• Configure a new connection:

  • JDBC driver class: net.ucanaccess.jdbc.UcanaccessDriver

  • JDBC URL: jdbc:ucanaccess://<filepath> , e.g. jdbc:ucanaccess:///home/johndoe/Documents/fooDB.accdb

  • In the “SQL Dialect” field, select “MySQL < 8”

• Create a new dataset using this connection to import an MS Access table in DSS.

Connecting to MS Access file with Date/Time data type

Be sure to follow the above steps, but utilize UCanAccess 4.0.4