Work with semantic views

A Snowflake semantic view and a Databricks metric view are governed metric layers your team defines in the warehouse. Golden can import one and chart its metrics live, the same way it queries a table — you never write SQL against it.

Import a semantic view

In the import wizard's Browse Tables tab, a semantic view or metric view appears as an ordinary row with a "Semantic view" or "Metric view" tag after its name. Tick it like any table and click Import Schema.

The Metrics tab in Prepare

When a dataset is only a semantic view, Prepare shows a Metrics tab in place of Table view. It lists the view's metrics under "Totals for the whole view", with each metric's aggregation shown beneath its name, and a "By" picker that breaks the metrics down by a dimension (the top 50 by the sort metric).

Chart a semantic view

The view's metrics appear under Measures and its dimensions under Dimensions, so you build charts the usual way. A semantic view has no rows of its own, so a chart needs a metric: drop only a dimension and Golden shows "Add a metric to draw this chart" and offers the view's metrics as one-click buttons.

• A fact can be grouped on but not measured.
• Only the view's metrics can be measured — a plain dimension can't.
• Per-record charts, such as a raw scatter or one row per record, aren't available; group by a dimension and chart a metric instead.

Calculated fields and joins

• A calculated field can combine the view's metrics with arithmetic and functions like ROUND, ABS, and IF — for example Revenue ÷ Orders. Dimensions and aggregates such as SUM aren't available, because the view holds no rows.
• You can join a semantic view with another table and chart across both. The relationship has to use one of the view's dimensions, not a metric, and filters aren't available yet on a chart that combines a semantic view with another table.