DQX Installation Wizard

When you install DQX as a tool in a Databricks workspace using databricks labs install dqx, the installer runs an interactive wizard to collect settings for the "default" run configuration. This page documents each installer prompt with its expected input, default value, and the behavior it controls.

When you use DQX's installation wizard to install DQX:

• Each prompt shows its default value in brackets. Press Enter to accept the default value.
• Some prompts will only appear conditionally (for example, the streaming, quarantine, metrics, and job-cluster prompts).
• When configuring locations (e.g. the input, output, or quarantine locations), the special value skipped instructs the installer not to configure the location.

Changing settings after installation

All responses map to fields in the generated configuration file. Add run configurations or change any of these settings by editing config.yml (open it with databricks labs dqx open-remote-config).

General settings

Configures the logging settings used across DQX jobs installed using the installer.

Prompt	Property	What it configures	Default	Notes
Log level	log_level	Logging verbosity for DQX workflows.	INFO	Accepts standard levels such as DEBUG, INFO, WARN, ERROR.

Input data

Defines an input_config where the source data is read. The input is optional during installation. Leave the location as skipped to omit it and configure it later.

Prompt	Property	What it configures	Default	Notes
Should the input data be read using streaming?	input_config.is_streaming	Whether the input is read as a streaming source rather than a batch source.	no	When enabled, additional streaming trigger options are requested for the output and quarantine tables.
Provide location for the input data	input_config.location	Source data, as a path or a catalog.schema.table / schema.table name.	skipped	Use skipped to omit the input configuration.
Provide format for the input data	input_config.format	Input data format, e.g. delta, parquet, csv, json.	delta	Only asked when an input location is provided.
Provide schema for the input data	input_config.schema	Optional explicit schema, e.g. col1 int, col2 string.	skipped	Only asked when an input location is provided. Use skipped to let the format infer the schema.
Provide additional options for reading the input data	input_config.options	Reader options as a JSON object, e.g. {"versionAsOf": "0"}.	{}	Only asked when an input location is provided.

Output data

Defines an output_config where checked data (with the _errors / _warnings reporting columns) is written. Leave the location as skipped to omit it and skip writing valid data (e.g. when you only want to write invalid rows to a quarantine table).

Prompt	Property	What it configures	Default	Notes
Provide output table	output_config.location	Output table, as catalog.schema.table / schema.table.	skipped	Use skipped to omit writing valid data. The output table can only be skipped if a quarantine table is provided.
Provide write mode for output table	output_config.mode	How results are written.	append	One of append or overwrite.
Provide format for the output data	output_config.format	Output data format.	delta
Provide additional options for writing the output data	output_config.options	Writer options as a JSON object, e.g. {"mergeSchema": "true"}.	{}
Provide additional options for writing the output data using streaming	output_config.trigger	Streaming trigger options, e.g. {"availableNow": true}.	{}	Only asked when streaming is enabled.

Quarantine data

Defines a quarantine_config where quarantined rows that fail 1 or more DQX checks are written. If the location is left as skipped, invalid rows are written to the output table instead.

Prompt	Property	What it configures	Default	Notes
Provide quarantined table	quarantine_config.location	Quarantine table, as catalog.schema.table / schema.table.	skipped	Use skipped to keep invalid rows in the output table. The remaining quarantine questions are then not asked.
Provide write mode for quarantine table	quarantine_config.mode	How quarantined rows are written.	append	Only asked when a quarantine table is provided. One of append or overwrite.
Provide format for the quarantine data	quarantine_config.format	Quarantine data format.	delta	Only asked when a quarantine table is provided.
Provide additional options for writing the quarantine data	quarantine_config.options	Writer options as a JSON object.	{}	Only asked when a quarantine table is provided.
Provide additional options for writing the quarantine data using streaming	quarantine_config.trigger	Streaming trigger options.	{}	Only asked when a quarantine table is provided and streaming is enabled.

Summary metrics

Defines the metrics_config and custom_metrics used to track and write per-run summary metrics produced by the quality checker.

Prompt	Property	What it configures	Default	Notes
Do you want to store summary metrics from data quality checking in a table?	metrics_config	Whether summary metrics are persisted.	no	When no, the remaining metrics questions are not asked.
Provide table for storing summary metrics	metrics_config.location	Metrics table, as catalog.schema.table / schema.table.	required	Only asked when storing summary metrics; must be provided.
Provide write mode for metrics table	metrics_config.mode	How metrics are written.	append	Only asked when storing summary metrics. One of append or overwrite.
Provide format for the metrics data	metrics_config.format	Metrics data format.	delta	Only asked when storing summary metrics.
Provide additional options for writing the metrics data	metrics_config.options	Writer options as a JSON object.	{}	Only asked when storing summary metrics.
Provide custom metrics	custom_metrics	Optional list of Spark SQL aggregate expressions to track, e.g. ["avg(salary) as avg_salary"].	[]	Only asked when storing summary metrics. Leave blank to track only the default data quality metrics.

Quality checks location

Defines a checks_location where quality check definitions are stored. The check definitions can be stored in a table or file.

Prompt	Property	What it configures	Default	Notes
Provide location of the quality checks definitions	checks_location	Where quality checks (rules) are stored.	checks.yml	Accepts a file name (relative to the installation folder), a catalog.schema.table / schema.table table, or a full /Volumes/.../<file> path.

Profiler

Defines a file path where the DQX Profiler writes summary statistics about the profiled datasets.

Prompt	Property	What it configures	Default	Notes
Provide filename for storing profile summary statistics	profiler_config.summary_stats_file	File produced by the profiler workflow.	profile_summary_stats.yml

Compute

Controls the compute used by the profiler, quality checker, and end-to-end workflows. Serverless is recommended; choosing job clusters unlocks per-workflow Spark configuration.

Prompt	Property	What it configures	Default	Notes
Do you want to use standard job clusters for the workflows execution (not Serverless)?	serverless_clusters	Compute type for the workflows.	no (use Serverless)	Answer no to keep Serverless (recommended). Answer yes to use job clusters, which triggers the per-workflow prompts below.
Optional spark conf to use with the profiler / data quality / end-to-end workflow	profiler_spark_conf, quality_checker_spark_conf, e2e_spark_conf	Per-workflow Spark configuration as a JSON object, e.g. {"spark.sql.ansi.enabled": "true"}.	{}	Only asked when not using Serverless. Asked once per workflow.
Optional Cluster ID to use for the profiler / data quality / end-to-end workflow	profiler_override_clusters, quality_checker_override_clusters, e2e_override_clusters	An existing cluster to reuse, e.g. {"default": "<existing-cluster-id>"}.	{}	Only asked when not using Serverless. If left empty, a job cluster is created automatically when the job runs.

Reference tables

Configures reference tables used by DQX checks (e.g. for schema validation or dataset comparison checks).

Prompt	Property	What it configures	Default	Notes
Provide reference tables to use for checks	reference_tables	Reference datasets for checks such as referential integrity, as a JSON map of name to an input specification.	{}	The specification accepts location, format, schema, options, and is_streaming. Example: {"reference_vendor": {"location": "catalog.schema.table", "format": "delta"}}.

Custom check functions

Configures a mapping used to reference custom check functions written in PySpark. Custom checks should be stored in your Databricks workspace as Python modules. Each key is a check function defined in the associated module.

Prompt	Property	What it configures	Default	Notes
Provide custom check functions	custom_check_functions	Custom check functions, as a JSON map of function name to a Python module path in the workspace or a volume.	{}	Example: {"my_func": "/Workspace/Shared/my_module.py"}.

Dashboard SQL warehouse

Configures a Databricks SQL warehouse for serving DQX's built-in dashboard.

Prompt	Property	What it configures	Default	Notes
Select PRO or SERVERLESS SQL warehouse to run data quality dashboards on	warehouse_id	SQL warehouse used by the quality dashboard.	select from list	Choose an existing PRO or SERVERLESS warehouse, or create a new PRO warehouse.

Dependencies

Configures installation of DQX from files in the workspace instead of PyPI. Useful for installing DQX jobs in air-gapped environments with no access to PyPI.

Prompt	Property	What it configures	Default	Notes
Does the given workspace block Internet access?	upload_dependencies	Whether DQX dependencies are uploaded to the workspace instead of being fetched from PyPI at runtime.	no	Answer yes for workspaces without Internet egress (air-gapped).