LookML in. Anything out.

LookML is code, which makes Looker estates easier to read than most. The hard part is everything around the code: which explores are actually used, what the Liquid templating does at query time, and how access filters combine with user attributes.

What we read

• Models and explores — joins, relationships and the fields each explore exposes.
• Views — dimensions, measures and dimension groups, including their SQL.
• Derived tables — SQL-based, native and persistent derived tables, and how they’re rebuilt.
• Dashboards and Looks — LookML dashboards and user-defined dashboards, tiles and filters.
• Access controls — access filters, sql_always_where and user attributes.
• Usage — which Looks, dashboards and explores people actually query.

How it translates

Where the care goes

• Symmetric aggregates. Looker protects measures from fan-out joins automatically. On the target, the model’s grain has to be right, and parity tests prove it.
• Liquid templating. Templated filters and parameters are resolved into explicit parameters or field selections.
• Dimension groups. Timeframes become a shared date table instead of per-view variants.

Moving to Looker

Looker is a target as well. From open specifications, our automated migration tool writes LookML views, explores and dashboards that your team can review like any other code.