Enabling Statistics and Profiling Features for Data Assets – DataHub Customer Support Portal

Area: Observability Issues

Sub-Area: Data Profiling and Smart Assertions

Issue

The Statistics tab appears greyed out on dataset pages in DataHub, preventing users from viewing profiling data such as row counts, null counts, min/max values, and other table statistics. This commonly occurs when profiling is not enabled in ingestion recipes or when Smart Data Quality assertions are not activated in the DataHub environment.

You Might Be Asking

Why is my Stats tab greyed out and disabled?
How do I enable data profiling for my datasets?
What configuration is needed to see statistics in DataHub?
How do I enable Smart Data Quality assertions?

Solution

The Stats tab is greyed out by design when there is no profiling data available. To enable it, you need to configure profiling in your ingestion recipes:

Add profiling configuration to your ingestion recipe:

source:
  type: databricks  # or your source type
  config:
    workspace_url: https://.azuredatabricks.net/
    # ... other config ...
    profiling:
      enabled: true
      method: "ge"  # or "analyze" for Delta tables only
      warehouse_id: ""

For other data sources, the profiling configuration varies:

# Snowflake example
profiling:
  enabled: true

# BigQuery example  
profiling:
  enabled: true
  profile_table_level_only: false

Save the updated recipe and re-run ingestion to collect profiling data for your datasets.
Verify the Stats tab becomes active after successful profiling ingestion.

For Smart Data Quality Assertions (DataHub Cloud)

If you need Smart DQ assertions enabled, contact DataHub Support to activate the feature flag in your environment:

Smart Assertions are a Cloud-only feature that requires backend configuration
Once enabled, you can create freshness, volume, schema, and field assertions via the UI
Navigate to a dataset → Validations tab → Create Assertion

Common Issues and Troubleshooting

Missing method field: Ensure your profiling block includes a valid method (e.g., "ge" or "analyze")
Wrong ingestion source: Verify you're updating the correct recipe if you have multiple sources
Permissions: Ensure your service principal has SELECT access to tables you want to profile
Large datasets: Use profile_pattern to limit which tables get profiled to manage run times

Additional Notes

Profiling adds queries against your data sources and will increase ingestion run time and compute usage. For environments with many tables, consider scoping profiling to critical datasets using allow/deny patterns. The Stats tab requires at least one of the following data types: table profiling data, partition profiling data, usage statistics, or operations history.