Skip to main content

Window Decorators

Moving averages, ranking, and window functions.

Overview​

Window decorators enable window functions like moving averages, running totals, ranking, and lead/lag calculations.

Available Decorators​

Moving Average​

Calculate moving average over a window.

# Applied in query, not in model definition
# User selects "Moving Average" decorator

Running Total​

Calculate running/cumulative total.

# Applied in query

Ranking​

Rank values (rank, dense_rank, percent_rank).

# Applied in query

Lead/Lag​

Access previous/next values.

# Applied in query

How It Works​

Window decorators are applied at query time:

  1. User selects measure and window decorator
  2. Planner generates SQL with window function
  3. Returns windowed values

Use Cases​

Moving Average​

- type: measure
  name: Revenue
  data_type: decimal
  expression:
    sql: sum(amount)

# User applies "7-day moving average" decorator
# Returns: 7-day moving average of revenue

Running Total​

# User applies "Running Total" decorator
# Returns: Cumulative sum over time

Ranking​

# User applies "Rank" decorator
# Returns: Ranked values (1, 2, 3, ...)

Best Practices​

  1. Use with date dimensions - Window functions often require ordering by date
  2. Document measures - Help users understand window options
  3. Test calculations - Verify window functions work correctly

Next Steps​