dbt Staging Model Lineage

Compiled SQL from a typical dbt staging → intermediate → mart layering. Column lineage shows source freshness from source tables to mart columns.

The lineage

The SQL

CREATE TABLE raw_app.users (
  id NUMBER,
  email STRING,
  signup_at TIMESTAMP,
  plan STRING
);

CREATE TABLE raw_app.events (
  event_id STRING,
  user_id NUMBER,
  event_name STRING,
  occurred_at TIMESTAMP
);

CREATE OR REPLACE VIEW analytics.stg_users AS
SELECT
  id AS user_id,
  email,
  signup_at,
  plan
FROM raw_app.users;

CREATE OR REPLACE VIEW analytics.stg_events AS
SELECT
  event_id,
  user_id,
  event_name,
  occurred_at
FROM raw_app.events;

CREATE OR REPLACE VIEW analytics.int_user_activity AS
SELECT
  u.user_id,
  u.plan,
  COUNT(e.event_id) AS event_count,
  MAX(e.occurred_at) AS last_event_at
FROM analytics.stg_users u
LEFT JOIN analytics.stg_events e ON e.user_id = u.user_id
GROUP BY u.user_id, u.plan;

CREATE OR REPLACE VIEW analytics.mart_active_users AS
SELECT user_id, plan, event_count, last_event_at
FROM analytics.int_user_activity
WHERE event_count >= 5;

What the lineage shows

• mart_active_users.event_count flows from raw_app.events.event_id via stg_events.
• mart_active_users.plan flows from raw_app.users.plan via stg_users.
• Visualizing compiled dbt SQL avoids needing to mentally re-walk ref() chains.