Data quality you can trust
DQX is the data quality framework for Databricks — define, monitor, and act on data quality issues across your lakehouse.
Provided by Databricks Labs
DQX Core
The Python package. Define and run checks in your PySpark pipelines and notebooks — batch and streaming.
Battle-tested data quality frameworkOpen the Core guideDQX StudioBeta
The no-code web app. Author, run, and monitor quality rules from your browser — no code required.
No code requiredOpen the Studio guideCompare what each option offers
DQX comes in two flavors. Pick one to see its features.
Data format agnostic
Works seamlessly with PySpark DataFrames.
Batch & streaming
Spark batch and Structured Streaming, with Lakeflow (DLT) pipeline integration.
Row & column-level rules
Define quality rules at both the row and column level.
Rich failure detail
Get detailed insight into exactly why a check failed.
Custom reactions to failures
Drop, mark, or quarantine invalid data flexibly.
Profiling & rule generation
Profile input data and auto-generate data quality rule candidates.
Code or config checks
Define checks in code, or declaratively as configuration.
Validation summary & dashboard
Track and pinpoint data quality issues over time.
Row anomaly detection
Detect unusual rows with trained ML models and explanations.
Data contracts
Generate quality rules from ODCS contracts, including schema validation.
MCP for AI coding agents
Expose DQX tools to Genie Code, Cursor, and Claude Code over MCP.
Actions and alerting
Send Slack, Teams, or webhook alerts, or fail the pipeline, when summary metrics cross a threshold.