Build a dbt staging model from a raw SQL script

One SQL query, pulling straight from the raw orders table. Three other reports are already asking to copy it.

BeginnerRead 4 minDoing: an hour6 steps

▶ Start

What you'll have at the end

One dbt staging model other models call with ref(), replacing the pasted raw SQL

Before you start

You need

A raw table already loading into the warehouse, a dbt project connected to it, and a SQL query pulling from that table pasted into more than one report.

Not covered

Making the model incremental instead of a full view rebuild, which only pays off once the source table gets too large to rebuild on every run.

Leans on

Wire the raw table into dbt

Register the raw table as a dbt source
Build the staging model from the source

Prove the model before anyone depends on it

Compile the model and preview its rows
Build the model in the warehouse

Retire the pasted SQL

Point every report at the new model
Commit the change and open it for review

Where to go next

Give dbt a dev target

Point dbt at a safe development target first if this build would otherwise run straight against the warehouse everyone else reads from.

Turn an expensive nightly full-refresh model into an incremental dbt model

Switch this model off a full rebuild once the source table gets too big to scan on every single run.

Checked 25 Aug 2026

Part of the dbt (Data Build Tool) cookbook