Free dbt Schema.yml Generator

Turn any CREATE TABLE into a complete dbt scaffold in one click. Paste a CREATE TABLE statement - from Snowflake, PostgreSQL, Redshift, BigQuery, or ANSI SQL - and get three artifacts: a schema.yml with inferred unique and not_null tests, a staging SQL model using the canonical with source, renamed pattern, and a sources.yml entry with freshness checks. No warehouse connection, no dbt codegen, no account required.

Why use a dbt schema generator?

Writing schema.yml files by hand is tedious and error-prone. dbt codegen does this against a live warehouse connection, but sometimes you don't have one yet - you're reviewing a CREATE TABLE in a pull request, designing a new staging layer from a data-contract doc, or bootstrapping a dbt project offline. This tool solves the offline case. Feed it any CREATE TABLE and get a schema.yml starting point in seconds.

What tests get inferred?

PRIMARY KEY columns -> both unique and not_null tests. PK enforces both invariants.
Columns named id -> both unique and not_null by convention - treated as surrogate keys.
NOT NULL columns (without PK) -> not_null test. DDL explicitly requires a value.
Nullable columns -> no auto-tests. Add accepted_values, relationships, or dbt_utils.expression_is_true manually based on business logic.

Generated artifacts explained

schema.yml

A dbt YAML doc with version: 2, a single models: entry, column list with descriptions and tests. Column descriptions are inferred from naming conventions (_id columns become foreign-key descriptions, _at columns become event timestamps, status/amount/email get intuitive descriptions). Always review and rewrite with business context - good column documentation is the single highest-ROI item in a dbt project.

Staging SQL model

A .sql file using the dbt style-guide pattern: with source as (...), renamed as (...) select * from renamed. The source CTE pulls from {{ source('...', '...') }}; the renamed CTE is where you add column casts, renames, or surrogate keys like {{ dbt_utils.generate_surrogate_key(['order_id']) }}.

sources.yml

A sources.yml entry with loaded_at_field and freshness thresholds (warn after 12 hours, error after 24 hours - tune to your SLA). Lets you run dbt source freshness to detect stale ingestion.

Where to place the generated files

Staging SQL -> models/staging/<source>/stg_<table>.sql
schema.yml -> models/staging/<source>/schema.yml (one file per source folder is the dbt style-guide convention)
sources.yml -> models/staging/<source>/sources.yml
Run dbt build --select stg_<table>+ to compile, test, and materialize the model.