A unified data governance solution that helps manage, protect, and discover data across your organization
Hey Anuj, it looks like you’ve hit a current gap in the Purview Data Quality engine: although Unity Catalog lets you see DATE and TIMESTAMP columns in the asset schema, the Data Quality profiler today only supports a limited set of data types and marks those time-based types as “Unsupported.” That prevents both profiling/rule definition and using them as watermarks for incremental scans.
Here’s how you can work around it today:
- Create a Unity Catalog view (or table) that converts your DATE/TIMESTAMP columns into a supported type • For example, in Databricks SQL:
• By casting to STRING or a numeric type, Purview DQ will recognize the columns as “supported.”CREATE OR REPLACE VIEW catalog.my_schema.my_table_for_purview AS SELECT colA, CAST(my_date_col AS STRING) AS my_date_col_str, CAST(my_timestamp_col AS BIGINT) AS my_ts_epoch -- or AS STRING FROM catalog.my_schema.my_original_table; - Point your Purview scan at the new view (or register it as a separate asset) • Run a full metadata & data quality scan against that view. • You’ll now see
my_date_col_strormy_ts_epochavailable for profiling and rule-based checks. - Configure incremental scans using the new “watermark” column • In your scan settings, select
my_ts_epoch(or the string representation) as the watermark for incremental filters. - Keep an eye on the Purview roadmap • Full native support for Unity Catalog DATE/TIMESTAMP in Data Quality is on the backlog. You can upvote or file a feature request via Azure feedback to accelerate it.
Hope that helps unblock you in the meantime!
Reference docs
• Supported data sources & file types in Purview scans
https://docs.microsoft.com/azure/purview/purview-connector-overview
• Configure file-type limitations & complex types
https://docs.microsoft.com/azure/purview/sources-and-scans#file-types-supported-for-scanning