> For the complete documentation index, see [llms.txt](https://help.daasity.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://help.daasity.com/core-concepts/ai-analyst/reading-a-views-health.md).

# Reading a view's health

What a view's status, governed flags, data-through date and left-out fields tell you before you rely on a number.

Every view carries a few signals that tell you how far to trust a number from it. Your AI sees them on every view and is asked to mention them when they matter. You can see them yourself by asking your AI: "List my views with their health."

## Status

Each view has one status. It describes what was checked when the view was built. It is not a comparison against your dashboards, so tie out any number that's going in front of leadership.

| Status                         | What it means                                                                                                                                                                              | Use it for                                       |
| ------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------ |
| **builds\_and\_returns\_data** | The view built and returns rows.                                                                                                                                                           | Normal use.                                      |
| **partial\_coverage**          | More than half of the view's fields have no data in your account.                                                                                                                          | The fields that do have data.                    |
| **stale**                      | The newest data in the view is older than expected: more than a week for most data, or about nine weeks for syndicated retail data such as SPINS, which publishes on a four-week calendar. | Historical questions only, and say so.           |
| **empty\_no\_data**            | The view exists, but its source has no rows in your account.                                                                                                                               | Nothing yet - usually a connector you don't use. |

## Built on a derived table

A separate flag shows whether a view is built on a physical table or on a derived table (a Looker derived table or SQL defined in the model). Views on a derived table can differ slightly from dashboards that read a stored copy of that table. Use them for rates, trends and changes, and tie out absolute numbers. See [How your views are built](https://help.daasity.com/core-concepts/ai-analyst/how-your-views-are-built.md).

## Governed and ungoverned metrics

Each metric is marked **governed** or **ungoverned**:

* **Governed** metrics read from Daasity's curated data models, with your valid-order rules, timezone handling and metric definitions built in.
* **Ungoverned** metrics read from a source Daasity doesn't curate, such as a sheet you upload or a raw connector feed.

A view can be healthy and still contain ungoverned metrics. For a number you'll report, look for both: a healthy view and governed metrics.

## Data through

Each view reports the newest date in its data. Check it before quoting a recent number: if a question asks about last week and the view's data ends two weeks ago, the answer can't include last week.

## Fields with no data

If a field has no values at all in your account, your AI is told not to filter or group by it. The view still answers questions about its other fields.

## Left out of the build

Each view lists the metrics and fields that couldn't be carried over from your model, with the reason. A missing field is almost always on that list, not removed on purpose. See [How your views are built](https://help.daasity.com/core-concepts/ai-analyst/how-your-views-are-built.md#what-doesnt-carry-over).

## Where answers come from

Your AI answers from these views first. When no view can answer, it may write SQL against Daasity's curated data models, and only after that against raw connector data. Any answer that didn't come from a view names the table it came from and says it should be reconciled against your dashboards. If you'd rather your workspace answer from views only, contact your account manager or <support@daasity.com>.
