Dictum - Metrics as code

How It Works

Define

Model your tables, metrics, and dimensions in readable syntax

 metric revenue = [orders:
sum(amount)
]
dimension channel = [orders: utm_source]

Query

Write natural queries with filters, groupings, and transforms

 filter channel where revenue.top(5)
select revenue.percent by channel

Execute

Get optimized SQL and results for any database

 -- Compiles to optimized SQL WITH ranked AS (
SELECT utm_source, revenue,
ROW_NUMBER() OVER (...)
  ...
)

Why Dictum

📝

Readable syntax

Not YAML, not raw SQL. A language designed for metrics that humans can read and review.

🔄

Version controlled

Definitions live in code. Review changes in PRs, track history, roll back mistakes.

⚡

Editor support

LSP integration for autocomplete, inline errors, and hover documentation.

🗄️

Multi-database

Works with DuckDB, ClickHouse, PySpark/Databricks. PostgreSQL, BigQuery, Snowflake coming soon.

🔗

Automatic joins

Define relationships once. Dictum resolves join paths automatically.

📊

Built-in transforms

.percent, .total, .top(N), .diff_percent — common analytics patterns built in.

Complete Example

# Define your metrics
metric revenue = [orders: sum(amount)]
metric orders = [orders: count(*)]
metric aov = revenue / orders  # derived metric

# Define dimensions
dimension channel = [orders: utm_source]
dimension country = [orders: customer.country]
dimension plan = [orders: customer.plan_name]

# Define table relationships
table orders
  joins customers as customer using (customer_id),
        products using (product_id)

# Query: Top channels by revenue with AOV
filter channel where revenue.top(5)
select revenue, revenue.percent, aov
by channel

with orders_75102368 as (
    select
        orders.utm_source as channel,
        sum(orders.amount) as revenue_75102368,
        count(*) as orders_75102368
    from orders as "orders"
    group by
        orders.utm_source
    order by sum(orders.amount) desc, orders.utm_source asc
    limit 5
),
orders_24912fa3 as (
    select
        orders.utm_source as channel,
        sum(orders.amount) as revenue_24912fa3
    from orders as "orders"
    group by
        orders.utm_source
),
merge_000c08a5 as (
    select
        orders_24912fa3.channel as channel,
        row_number() over (order by orders_24912fa3.revenue_24912fa3 desc, orders_24912fa3.channel asc) <= 5 as revenue
    from orders_24912fa3 as "orders_24912fa3"
),
orders_a0ee069b as (
    select
        sum(orders.amount) as revenue_a0ee069b
    from orders as "orders"
    where exists (select 1 from merge_000c08a5 where merge_000c08a5.revenue and merge_000c08a5.channel = orders.utm_source)
)
select
    coalesce(orders_75102368.channel, orders_75102368.channel, orders_75102368.channel, orders_75102368.channel) as channel,
    orders_75102368.revenue_75102368 as revenue,
    orders_75102368.revenue_75102368 / orders_a0ee069b.revenue_a0ee069b as revenue__percent,
    orders_75102368.revenue_75102368 / orders_75102368.orders_75102368 as aov
from orders_75102368 as "orders_75102368"
left join orders_a0ee069b as orders_a0ee069b
    on true
order by channel

Metrics as code.