Quality Rules
Quality is one of the three tabs in the area below the canvas. It runs declarative validation rules against the model you have open and reports, per rule, whether it passed, failed, or errored — with the offending records listed underneath.
Open it with Panels in the status bar, then the Quality tab.
Where the rules come from
Rules are not built into the client. They arrive as rule packs, installed per user from the Quality rules group in the Marketplace. Adding a rule or changing one is a content change: the pack is updated, you reinstall it, and nothing has to be rebuilt or redeployed.
That is also why the tab can be empty on a fresh deployment. No rule groups available means no pack is installed, not that the model is clean.
Running rules
| Control | What it does |
|---|---|
| Run all | every rule of every installed group |
| Run group | one group — referential integrity, custom properties, model hygiene, whatever the pack defines |
| Run rule | a single rule, which is what you want while fixing one thing |
| Enable rule | takes a rule out of the run without uninstalling the pack |
While a run is going you see progress per rule, not just a spinner.
Reading the report
Each rule ends in one of three states, shown as chips at the top: Passed, Failed, Errors. The distinction between the last two matters — failed means the rule ran and the model violates it; errors means the rule itself could not run, which is a problem with the pack, not with your model.
Expand a failed rule to see its violation rows. Show filters the report to All, Failed or Errors, which is the fastest way through a long report.
How it works, and what it cannot touch
A run projects the current model into a read-only in-browser database and evaluates the rules as SQL against it. Nothing is sent anywhere and nothing is written back: a quality run cannot change your model. The same projection backs the SQL tab, so anything a rule can find, you can also go looking for by hand.