# DQX Data Quality

> Open-source Databricks Labs framework to validate PySpark DataFrames and tables, with checks generated from profiling and defined in YAML or Python.

Source: https://toolsfordata.com/tools/dqx/ · Updated 2026-11-13

[Docs](https://databrickslabs.github.io/dqx/) · [GitHub](https://github.com/databrickslabs/dqx) (470+ stars)

Best for: Databricks teams validating batch and streaming pipelines who are fine with a Labs project that has no formal Databricks support SLAs.

## What it's used for

DQX is used to validate PySpark DataFrames and tables on Databricks, in batch or streaming: checks written in YAML or Python drop, mark or quarantine invalid rows, and the profiler suggests the first rules.

## Pricing

DQX is free to use with Databricks services under the Databricks License; you pay for the Databricks compute it runs on. As a Databricks Labs project, it has no formal support SLAs.

## Pricing

- Open source: OSS library
- Source: https://github.com/databrickslabs/dqx (checked 2026-11-02)

## Where it fits

- data testing (core): Open-source Databricks Labs framework where data tests are row- and column-level checks in YAML or Python on PySpark DataFrames, in batch, streaming, or Lakeflow pipelines, that drop, mark, or quarantine invalid rows.
- open source (core): Source-available data quality framework under the Databricks License, from Databricks Labs, free to use with Databricks services, with checks, quarantine, profiling, and anomaly detection for PySpark.
- data profiling (strong): Open-source Databricks Labs framework whose data profiler computes summary statistics on PySpark DataFrames and tables and generates candidate data quality rules from the data profile.
- data observability (partial): Open-source Databricks Labs framework with ML row anomaly detection that explains which columns drove each score, summary metrics tracked over time, and a quality dashboard, as data observability on Databricks.

## Integrations

2 integrations, each sourced from the vendor docs.

- Data Transformation: Apache Spark
- Data Warehouse: Databricks SQL

## Alternatives

- [Databricks Data Quality](https://toolsfordata.com/tools/databricks-data-quality/): Databricks' supported option: expectations in Lakeflow pipelines and Unity Catalog monitoring.
- [Spark Expectations](https://toolsfordata.com/tools/spark-expectations/): Consider Spark Expectations if bad rows should be quarantined before they reach the target table.
- [Deequ](https://toolsfordata.com/tools/deequ/): Consider Deequ if your data is large and processed in Spark.
- [SparkDQ](https://toolsfordata.com/tools/sparkdq/): Consider SparkDQ for a lighter, config-driven PySpark check library.
- [Great Expectations](https://toolsfordata.com/tools/great-expectations/): Consider Great Expectations, the best-known open-source framework, for Python checks on files, SQL databases and DataFrames.
