Supported data sources in Power BI

Power BI is business intelligence software that can integrate with various data sources. When you ingest Power BI metadata, Collibra Data Lineage tries to automatically stitch this metadata to data sources registered in Data Catalog. It also creates a technical lineage that shows where metadata is used and how it transforms.

The following table shows the supported data source types in Power BI that have been tested. If a data source is identified as certified, it means that the data source:

  • Is ingested in Data Catalog as a Power BI Data Model asset.
  • Is shown in the technical lineage and stitching is possible.
Tip When selecting your data sources in Power BI, the data source types under the following tabs require the specified connection types:
  • The Files tab refers to Parquet files. Connect via the URL to the files.
  • All data source types under Database require a JDBC connection.
  • For Power Platform, you need a live connection.
  • The three data source types under the Azure tab require either a JDBC or ODBC connection.
  • All data source types under Other require an ODBC connection.

The connections types are mentioned in the following table, for each supported data source type.
Important 
  • Although the following data sources have been tested extensively, there still may be some issues caused by unsupported elements within the data source or limitations in the Power BI integration process.
  • Collibra Data Lineage can connect only to datasets that are hosted by Power BI. It cannot connect to externally hosted datasets or models. For complete information, consult Microsoft's Power BI documentation.
  • Real-time datasets, semantic models with object-level security, semantic models with a live connection to AS-Azure and AS on-premises, and Excel full fidelity datasets aren't supported for subartifact metadata. For complete information, consult Microsoft's Real-time streaming in Power BI documentation.
  • SharePoint files are not a supported data source type. However, there is an Ideation ticket for this. If you'd like to vote for support for SharePoint files as a supported data source in Power BI, you can do so in the Collibra Ideation Portal.

Power BI data source

Connection type

Certified ?

Amazon Redshift

JDBC

Yes

Apache Hive

ODBC

Yes

Azure Analysis Services

ODBC

Yes

JDBC

No

  • Via import, technical lineage is possible only from the data set to the report. Stitching is not supported.
  • Via direct connection, technical lineage and stitching are not supported.

Azure Databricks

JDBC, ODBC

Yes

Collibra Data Lineage supports the following functions:

  • Databricks.Catalogs
  • Databricks.Contents
  • Databricks.Query
    Note For Databricks.Query, the ingestion success rate is high, but it's not 100%.
Azure Synapse Analytics

JDBC

Yes

Dremio

JDBC

No

Dremio data sources are ingested as Power BI Data Model assets and shown in the technical lineage, but stitching is not possible.

Google BigQuery

JDBC

Yes

Impala

ODBC

Yes

MySQL

JDBC

No

MySQL data sources are ingested as Power BI Data Model assets, but not shown in the technical lineage and stitching is not possible.

Netezza

ODBC

Yes

ODBC

ODBC

Yes

Important You need to use a Power BI <source ID> configuration file to provide the true system names of the ODBC databases in Power BI. For more information, see Providing ODBC database names in Power BI.

Oracle

JDBC, ODBC

Yes

Note If you connect via ODBC:
  • Oracle Views are supported.
  • In most cases, you need to use a Power BI <source ID> configuration file for database mapping, as the database name returned by the API differs from the true database name.

Parquet file

URL to the files

No

SAP HANA

ODBC

Yes

Snowflake

JDBC

Yes

SQL Server

JDBC

Yes

Sybase

JDBC

Yes

Power BI with measures only, no columns

JDBC

Yes

Power Platform > Power BI data set

Live connection

Yes

Important Supported only if the data set is from one of these supported data sources.

Note We cannot guarantee that other data sources in Power BI can be stitched successfully.