`weaver test`¶

Run Lakehouse Python/Spark SQL, Warehouse SQL and semantic-model DAX validations through one Test operation. Installed results share the catalogue and appear together in Health.

Synopsis¶

weaver test [-h] [--item ITEM] [--name PATTERN] [--file PATH] [--source SOURCE] [--concurrency N] [--dry-run] [--json] [--non-interactive] [--workspace WORKSPACE] [--workspace-config WORKSPACE_CONFIG] [--environment ENVIRONMENT] [--catalogue CATALOGUE] [ITEM ...]

Positional arguments¶

ITEM: Weaver items to validate, as Lakehouse/Name, Warehouse/Name or SemanticModel/Name. PowerBI or PowerBI/ names the semantic models of the project folder. Naming none validates every item.

Options¶

-h, --help: show this help message and exit
--item ITEM: Legacy spelling for a positional Weaver item.
--name PATTERN: Run only validations whose whole Schema.Object matches this regular expression, with their diagnostic rows. Repeat to select more.
--file PATH: Run validations from source in file mode: a file, a directory or a glob. A file outside the project folder runs against the one item named. Repeat to select more.
--source SOURCE: Project folder, or an abfss location inside a Fabric session. Defaults to the current directory or Notebook Resources.
--concurrency N: Limit total simultaneous execution for this invocation; retain resource limits.
--dry-run: Show the test plan without running it.
--json: Emit the report as JSON.
--non-interactive: Do not read stdin, wait for a keypress or open browser sign-in. Missing authorisation or required input is an error.
--workspace WORKSPACE: Fabric Workspace name.
--workspace-config WORKSPACE_CONFIG: Workspace configuration file.
--environment ENVIRONMENT: Fabric Environment name or Workspace/Environment reference.
--catalogue CATALOGUE: Where the Weaver catalogue lives, for example Warehouse/Weaver.

Responsibility and selection¶

Test evaluates Tests/Assumptions and returns a report. Positional/legacy items define scope. With a catalogue and no files it uses installed definitions. --file selects source/file mode even with a catalogue; without a catalogue Test reads project validations. Source/file mode records no installed TestStatus/Log. Reports cannot Test; validate their model.

Repeat --name for whole-name case-insensitive regex filtering; every pattern must match. Repeat --file for source files, project directories or quoted glob patterns. The selectors can be combined. Project-declared files keep their owners; an external single file requires exactly one named item. --source supplies project context (default current directory/Notebook Resources); it alone does not force file mode when a catalogue exists. PowerBI/project selectors expand semantic models.

Supported source forms include Warehouse SQL, Lakehouse Python/Spark SQL and model-owned DAX. DAX reconciliation still requires its Expected Lakehouse/Warehouse and query permissions. Name/file-targeted runs collect diagnostics; item-wide runs normally report counts. --dry-run parses/resolves without evaluating queries or recording state.

Execution and outcomes¶

Validations are independent: findings and node evaluation errors do not prevent later selected validations. Installed runs use configured resource limits. Source/file mode is serial by default; explicit --concurrency N uses scheduled execution with a total cap alongside resource ceilings. It does not Load data dependencies or refresh a model first.

No edit-and-retry loop is offered. Non-interactive mode disables prompts/browser sign-in. Human output/JSON distinguish passed, failed, invalid, planned, with missing/unexpected Test counts, Assumption violations and available diagnostic rows. Python returns completed reports including findings; weaver.test has no strict-mode argument. CLI exits 0 for passed/planned, 1 for failed/invalid or command error, 2 for parser usage errors.

Examples¶

weaver test Warehouse/Operations --name 'Parcel[.]StatusMatches' --workspace-config workspace-dev.yml
weaver test Warehouse/Operations --file Warehouse/Operations/tests --source . --concurrency 2 --workspace-config workspace-dev.yml
weaver test SemanticModel/Parcel --source . --workspace-config workspace-dev.yml

The last command uses file mode only when that configuration has no catalogue; with a catalogue it selects installed validations. Add --file PowerBI/Parcel/assumptions/Parcel to explicitly choose source mode.

See Test behaviour, DAX validations and machine output.

Validate source data and semantic models together¶

With the source and model validations installed:

weaver test --workspace-config workspace-config.yml

The same report contains Lakehouse/Warehouse validations and model-owned DAX validations. A DAX Test can compare a measure with SQL calculated independently over its managed source. To inspect one model, select SemanticModel/Parcel; Reports have no separate Test. See Validate a semantic model and Compare DAX with source SQL.

weaver test¶