Discover, connect, and explore data in Synapse using Azure Purview

In this document, you will learn the type of interactions that you can perform when registering an Azure Purview Account into Synapse.


Using Azure Purview in Synapse

The use Azure Purview in Synapse requires you to have access to that Purview account. Synapse passes-through your Purview permission. As an example, if you have a curator permission role, you will be able to edit metadata scanned by Azure Purview.

Data discovery: search datasets

To discover data registered and scanned by Azure Purview, you can use the Search bar at the top center of Synapse workspace. Make sure that you select Azure Purview to search for all of your organization data.

Search for Azure Purview assets

Azure Purview actions

Here is a list of the Azure Purview features that are available in Synapse:

  • Overview of the metadata
  • View and edit schema of the metadata with classifications, glossary terms, data types, and descriptions
  • View lineage to understand dependencies and do impact analysis. For more information about, see lineage
  • View and edit Contacts to know who is an owner or expert over a dataset
  • Related to understand the hierarchical dependencies of a specific dataset. This experience is helpful to browse through data hierarchy.

Actions that you can perform over datasets with Synapse resources

Connect data to Synapse

  • You can create a new linked service to Synapse. That action will be required to copy data to Synapse or have them in your data hub (for supported data sources like ADLSg2)
  • For objects like files, folders, or tables, you can directly create a new integration dataset and leverage an existing linked service if already created

We are not yet able to infer if there is an existing linked service or integration dataset.

 Develop in Synapse

There are three actions that you can perform: New SQL Script, New Notebook, and New Data Flow.

With New SQL Script, depending on the type of support, you can:

  • View the top 100 rows in order to understand the shape of the data.
  • Create an external table from Synapse SQL database
  • Load the data into a Synapse SQL database

With New notebook, you can:

  • Load data into a Spark DataFrame
  • Create a Spark Table (if you do that over Parquet format, it also creates a serverless SQL pool table).

With New data flow, you can create an integration dataset that can be used a source in a data flow pipeline. Data flow is a no-code developer capability to perform data transformation. For more information about using data flow in Synapse.

 Next steps