Test cases

Filtering and searching

The toolbar's Search box matches two fields only: the test case description and its SQL. Several search terms can be given, and a test case has to match every one of them. Everything else is filtered from its own column header:

• Connections/Test suites - filter by connections or test suites (with a "not set" option)

• Status - filter by test case latest execution status

• Severity, Tags, Labels - filter by the test case's properties

• Description, Expected, Result, Relation - text filters, each with a "not set" option. The Relation filter matches the test's schema, object or column, and also the combined schema.object form the column displays

• Execution time - filter by a specific date

• Custom fields - each custom field has its own column and filter

Applied filters show as a chip row that can be cleared in one click, restored from the filter history, or shared with other users via a copied link. The visible columns and their order can be customized from the table settings in the table header, which also holds the Row status colors display toggle.

Bulk actions

Selecting test cases with the row checkboxes (Shift+Click selects ranges; the header checkbox selects everything matching the current filters) replaces the toolbar with the bulk actions:

• Execute - Execute selected tests

• Assign - Edit one field on all selected tests at once: Test suite, Connection, Severity, Expected result, Business rules, Threshold, Enable/disable or a Custom field

• Delete - Delete selected tests

• More options - Export selected tests into JSON, Excel, ZIP or CSV file, or Duplicate them

• Cancel - Clear the selection

Test cases can be imported via More → Import tests as JSON, Excel or CSV files:

As-of date filter

The clock button in the toolbar rewinds the whole page to a chosen date and time: statuses, results, changes and execution histories all show what they were at that moment. The chosen time is kept in the page URL, so an as-of view can be shared as a link. Executing tests is disabled while the filter is active.

See the date filter for details.

Standard validations

Standard validations provide a fast way to create test cases from predefined validation types with minimal user input. Instead of writing SQL manually, users select a validation type and fill in a few details.

1. Click "New test case" button and select "Standard validations"

2. Choose a validation type from the list - the selected type expands into its form

   

3. After selecting a validation type, fill in the required details:

   • Connection - Target connection

   • Schema - Database schema

   • Object - Table or view

   • Column - Target column (for column-level validations)

   • Additional parameters depending on the validation type (e.g. allowed values)

4. The generated SQL and a live Expected / Actual readout are shown before saving

5. Click "Create" (or "Create and execute") to save the test case

Test cases can also be created as standard validations directly from the catalog object view.

Create new test case

The test case type of an existing test can be switched between Query, Query + total and Compare from the type selector in the sidebar.

SQL editor

The SQL editor on the SQL tab comes with several companions, reachable from the buttons in the editor's top right corner:

• Catalog browser - a searchable side panel with the connection's schemas, objects and columns, for browsing the metadata while writing the query

• Typeahead - the editor suggests schema, object and column names from the selected connection's metadata as you type, plus variables. The Autocomplete settings gear controls it: SQL suggestions (turn all suggestions off), Suggest SQL keywords and Auto-format SQL on open

• Preview run - the play button executes the query and shows the actual value with a pass/fail verdict against the expected result before saving; View rows opens the row-level results

Test case detailed view

Clicking a test case opens its detailed view. The sidebar holds the editable properties described above; the main area is organized into the Executions, SQL, Dependencies and Comments tabs. From the view's header the user can:

• Copy link - copy a direct link to the test case

• History - see who created and last modified the test case; with the View change history privilege, browse and restore full audit history versions

• Notifications - subscribe to alerts for this test case

• Schedule - see and set the test case's schedule

Two more toggles moved into the view itself:

• Disabled - the slider in the sidebar's Execution section excludes the test case from test suite executions

• Save results - the slider above the results pane saves sample results of executions as CSV files (automatically only for failing executions; a test suite that includes results in its report forces this on)

Test cases managed by dbt or belonging to a read-only test suite open read-only, marked with a pill in the header.

Test case scheduling

Test cases can be scheduled to run individually without having to be assigned to a test suite. This is done via the Schedule segment in the detailed view's header - its label always shows the current schedule (or "Not scheduled"):

If the test case's suite is scheduled, the segment shows the inherited suite schedule; a warning is shown when a schedule exists but the test case is disabled.

You can check scheduling specifics from scheduling

Generating SQL

DQM has rule templates, which can be used to generate SQL with just a few clicks.

1. Open a test and go to the SQL tab

2. Press the "Generate" button and select "Dynamic rules" (the menu also offers the AI Assistant and Standard validations)

   

3. Edit parameters:

   

   1. Connection - Connection to generate SQL for

   2. Parameters - Schema, object and columns selection to generate SQL for

   3. Rule selection - Predefined SQL rule template for SQL generation, with a live SQL preview

   4. SQL auto-update - Enable for updating SQL query based on metadata changes (requires all columns to be selected)

4. Press " Generate SQL "

Test case previous result

In addition to numerical and string values, users can utilize the "previous" functionality. This feature compares the current execution result with the result from the last execution. Type previous as the expected result's value:

Test case dqm_query_result

Normally, the query result is taken from the COUNT(*) projection. However, users can include additional projection items (or subquery) and assign the alias dqm_query_result to the one they want as the primary result.

Sample test case SQL Queries:

Test case comments

Test case comments are meant for collaboration between DQM users. Users can add and edit comments on test cases they have access to.

The Comments tab in the detailed view shows the comment count and an unread indicator. Comments are grouped by day; press Enter to send (Shift+Enter for a new line).

• Editing comments - Hover a comment to edit or delete it (Non-admin users can only edit their own comments)

• Adding new comment - Add new comment to test case

Dependencies

The Dependencies tab of the detailed view shows all upstream and downstream dependencies of the object used by this test case. When the test case has no relation set, the tab asks for one - set the schema and object in the Relation field first.

A selector switches the tab between Objects and Attributes, and two summary pills give the Downstream and Upstream object counts with their combined quality. Each row shows the dependency's Type, the object with its connection and schema, how many of its columns are involved, and its Quality.

Each dependency row has a Lineage action, counting the relations behind it, that opens the object- or column-level Lineage view.

Test case relations

Test case relations are automatically assigned for test cases that are generated from rule templates. Users can also manually assign relations to test cases.

The Relation field in the detailed view's sidebar shows the current relation (marked with an "Auto" badge when derived automatically). Opening it allows toggling the Automatic relation switch or picking the schema, object and columns manually; changes are saved with the test case.

Labels

Test case labels can be used to group test cases. These labels can be used for filtering tests or even executing certain test via DQM API. Labels are added in the Labels field of the detailed view's sidebar:

Execution graph

The Executions tab shows all executions for the test case alongside a prediction graph.

On execution graph you can see:

1. Executions - Time series data for latest executions. Clicking an execution in the history list highlights its point on the chart and loads its results

2. Prediction graph - Prediction graph for execution result history and allowed error rate

The pencil button above the execution history switches it into edit mode, where selected executions can be enabled, disabled (excluded from analysis) or deleted. The chart is shown for numeric results only.

Test case results

Count based test cases and compare test cases have test case results, which can instantly show users related results from query. The results are shown below the execution graph on the Executions tab.

Results will be displayed as a table with per-column filters, where users can copy the query, download results as CSV or Excel, and expand the table to full screen.

When "Save results" is enabled, every execution has a persisted results file, and the diff pills (New/Resolved/Active) show how the failing rows changed.

Results will be displayed as two comparison tables, where users can see which records are the closest match and export both results to CSV files.