LookML puts join structure and query content in one reusable semantic model

LookML is the language Looker uses to create semantic data models. Analysts describe dimensions, aggregates, calculations, and table relationships; Looker builds SQL against a chosen database. The docs cast LookML as a dependency language like make, with predefined types and syntax, independent of SQL dialect, and usable without prior programming experience. Projects hold model files (tables and joins) and view files (how fields are calculated), typically under Git.

For data teams, the DRY pattern matters: SQL expressions live in one place and are reused for ad hoc queries. Structure (joins) is separated from content (columns, derived fields, aggregates, filters), so business users can explore via Explore without learning LookML or SQL.

That split is a practical form of governed, shared meaning: the semantic model is the contract between warehouse tables and the questions people ask, kept as versioned files that both humans and the SQL generator read.