Transformations
SQL that derives a new dataset, the method stated once, re-run when its inputs change.
A transformation is a saved query that derives a new dataset from others: a driver adapted to a model's grain, a join that feeds a model, a mapping applied. The query is the method, written down once; the notes say what was done and why.
Making one
On Data, choose New → Transformation. Name the output, write the query over your datasets (their names autocomplete), Preview the first rows, then Create transformation. On its page, click Run to produce the output.
The transformation and its output are one page: it opens on the result rows, and Edit shows the query and the output's columns. Edits save as you type, but the output changes only when you Run it.
When the inputs change
When a dataset it reads changes, the transformation's page says so — inputs changed since the last run — and offers Run to bring the output up to date. Charts and views that read the output flag it as out of date until you do.
To bring a whole project up to date at once, open Lineage from the project's Overview and click Run: every transformation runs in order, upstream first.
What to use it for
Deriving, not cleaning. Aggregate a driver weighted by its base, allocate a total by a sourced key, interpolate between survey years, join the drivers a model needs into one feed. Each of these is a method worth naming, and a transformation names it.
Faster with an agent — describe the table you want: "adapt the adoption survey to yearly values, interpolating between survey years, and say so in the notes." Your agent writes the query, runs it, documents the method, and you review the result.