Share via


Data governance known issues

This article provides information about known data governance issues in Microsoft Purview Unified Catalog management, discovery, and health management.

Unified Catalog management issues

Issue ID  Description Publish Date  Status  Fixed Date 
Data assets from top-level hierarchies aren’t selectable.

While adding data assets to data products, assets that are part of a semantic model or PBI Dataset hierarchy are automatically expanded without getting user input on which ones to include. 
9/6/2024  Active 
Missing duplication checks.

All business concepts except governance domains are missing duplication checks during creation. 
9/6/2024  Active 
Incomplete workflow then Privacy reviewer is selected.

Privacy reviewer selection skips the last step of showing data assets table for recording asset level access provisioning status and instructions to data consumer. 
9/6/2024  Active 
Inconsistent asset details page experience.

In the asset discovery experience under Discovery->Data assets, the asset details page for assets within the default/merged account render in the new Microsoft Purview portal but assets within accounts that were merged to the default account render in the classic governance portal. 
10/25/2024  Active 

Discovery issues

Issue ID  Title  Publish Date  Status  Fixed Date 
Enterprise glossary filters don’t interact.

Filters in the enterprise glossary don't automatically update based on other filter selections.  
9/6/2024  Active  9/26/2024
Data asset metadata don’t autorefresh.

Data assets within data products don’t automatically reflect owner, description, or lineage updates made in the data map. They need to be manually refreshed in the Catalog Management->Data products section by clicking on the Refresh button in the asset details page. 
9/6/2024  Active 
Linked OKRs don't show in data product search.

Published OKRs that are linked to data products, don't display in the data product details page within the Discovery -> Data products search experience today. 
9/6/2024  Active 

Health management issues

Issue ID  Title  Publish Date  Status  Fixed Date 
Virtual network isn't supported for Fabric and any native connectors including Google BigQuery, Databricks Unity Catalog, and Snowflake.  9/1/2024  Active 
No SQL query is generated for custom Data Quality (DQ) rules failures. SQL queries are only generated for failures of out-of-the-box rules.  4/11/2024  Active 
Data quality profiling uses a random sample of 1 million records. Users can include a maximum of 50 columns in a single batch, but they can run multiple batches if they need to profile files or tables with more than 50 columns.  4/11/2024  Active  No plan to fix this as this isn't a blocker.
A table lookup rule can be created for the tables within the same governance domain. You need to use four characters when you search for a data asset to reference to look up.  4/11/2024  Active 
The freshness Data Quality rule isn't supported for Azure SQL, Google BigQuery, Databricks Unity Catalog, Snowflake, and Synapse.  9/1/2024  Active 
Azure SQL doesn't support RLIKE keywords, so any query containing RLIKE keywords won't execute in Azure SQL. Custom regex rule won't work for Azure SQL.  4/11/2024  Active 
Fabric OneLake tables and semantic model data quality scanning isn't supported. Only Fabric Lakehouse tables data quality scanning is supported (private preview).  9/1/2024  Active 
Scheduled jobs have a maximum limit of 30 data assets. Customers need to set up multiple schedules if they have more than 30 data assets in a governance domain.  9/1/2024  Active 
9 Data Quality threshold alerts can be set up for a maximum of 30 data assets in one batch. If customers have more than 30 data assets in the same governance domain, they need to create multiple alerts, for example: Alert 1 for first 30 data assets and Alert 2 for next 30 data asset, and so on. 9/1/2024  Active 
10  Reports show all records (metadata), no filter available for role-based access control (RBAC). This means a data steward from one org (part of collection 1) can view analytics and reports from another org (collection 2). Recommending to use Purview metadata from OneLake to create own reports and insights. 9/1/2024  Active 
11  Not support N arbitrary Hierarchies of Directories with Parquet Files. We advise the customer to present data in
(1) A directory with Parquet Part File. For example: ./Sales/{Parquet Part Files} or
(2) A directory with Partitioned Parquet Files, partitioned by columns within the dataset like sales data partitioned by year and month. For example: ./Sales/{Year=2018}/{Month=Dec}/{Parquet Part Files}.
9/1/2024  Active