Supported data sources for data source registration

Collibra Data Intelligence Cloud supports several databases to register as a data source. Depending on your data source, you can use Collibra-provided Catalog connector, or your own JDBC driver.

Your own JDBC drivers

For certain data sources, you can use your own JDBC driver. The following table contains the packaged data sources and versions that have been tested.

Data source

Tested versions

Support for profiling and sample data

JDBC driver version

Amazon Redshift

1.0

Yes

v. 1.1.13.1013

Cloudera Hive

5.10 - 5.14

No

Apache driver v. 1.2.1

Hortonworks Hive

2.5, 2.6

No

Apache driver v. 1.2.1

HP Vertica 7.0

Yes

v. 07.01.0200

IBM DB2 10.5

Yes

v. 4.9.78

MySQL 5.6, 5.7

Yes

v. 5.1.38

Oracle 11g, 12c

Yes

v. 12c

PostgreSQL 9.4, 9.5

Yes

v. 9.4.1207

Microsoft SQL Server

2014, 2016

Yes

v. 5.1.38

Note Only Microsoft drivers and drivers available via Collibra Marketplace are supported.

Teradata 15.0, 16.20.07.01

Yes

No driver tested

Snowflake      

Note We cannot guarantee that other data sources or driver versions work correctly. If you use a generic JDBC driver or an unsupported version, data ingestion, data profiling and sample data may not work as expected.

Authentication and permissions

Both ingestion andprofiling(including sampling and advanced data type detection) rely on JDBC drivers to operate. Those drivers authenticate to the data sources as a user registered in that data source with specific permissions attached to the user profile in the data source.

To ingest a database without profiling, Data Catalog requires read access to the database metadata: description of schema, tables, columns, including some more complex properties such as the primary and foreign keys.

However, if you enable one or more profiling options, Data Catalog also requires the permission to read the full table. Which permissions are required exactly depends on the data source type, version and configuration. Additionally, they can also differ according to the provider and version of the JDBC driver. Most of the queries required to retrieve the information above are hidden by the driver. As a consequence, Collibra cannot give a exhaustive list.

Note Collibra supports several authentication methods, including credentials, NTLM, CyberArk and Kerberos. If you are using a certified Collibra provided driver on the CollibraMarketplace, you can also authenticate using Windows Authentication.

Tip If you need more detailed information, we recommend to contact your JDBC driver provider.

Warning Currently, if you register a data source viaEdge, you cannot profile your metadata and you can only use username/password authentication.