SQL Models#

Write models in pure SQL. Interlace automatically detects dependencies from your table references.

Basic SQL Model#

Create a .sql file in your models directory:

-- models/active_users.sql
SELECT *
FROM users
WHERE status = 'active'

Referencing Other Models#

Interlace parses your SQL and automatically detects dependencies from FROM and JOIN clauses:

SELECT
    o.id,
    o.amount,
    c.name AS customer_name
FROM orders o
JOIN customers c ON o.customer_id = c.id

Dependencies are extracted automatically—no special syntax required. Simply reference other models by their table name.

Model Configuration#

Add configuration in a YAML block comment:

-- models/daily_metrics.sql
/*
interlace:
  materialise: table
  strategy: replace
*/

SELECT
    DATE_TRUNC('day', created_at) AS date,
    COUNT(*) AS order_count,
    SUM(amount) AS total_amount
FROM orders
GROUP BY 1

Available Configuration Options#

When to Use SQL#

SQL models are ideal for:

• Complex queries that are hard to express in Python/ibis
• Porting existing dbt or SQLMesh models
• Team members more comfortable with SQL
• Database-specific optimizations and functions
• Simple transformations that don’t need Python logic

Mixing Python and SQL#

You can freely mix Python and SQL models in the same project. A SQL model can depend on a Python model and vice versa:

# models/enriched.py
@model(name="enriched_orders", materialise="table")
def enriched_orders(daily_metrics: ibis.Table, customers: ibis.Table) -> ibis.Table:
    # daily_metrics comes from a SQL model
    return daily_metrics.join(customers, ...)