dimensionalize (previously exact_grains)

info

We have renamed exact_grains to dimensionalize to align with how the function is used in practice. While we will continue to support exact_grains for backward compatibility, please use dimensionalize to ensure that your code is readable.

Definition

Evaluate an expression using a specific Level of Detail (LoD) context, regardless of the outer context. You can use dimensionalize for use cases like dimensionalized measure and cohort analysis.

caution

This is only allowed in the definition of a Dimension, and not elsewhere.

Syntax

dimensionalize(measure, dimension, ...)

Examples

dimension order_value {
  description: "Total order value of this user"
  definition: @aql dimensionalize(sum(order_items.order_value), users.id) ;;
}

// with pipe
dimension order_value {
  description: "Total order value of this user"
  definition: @aql sum(order_items.order_value) | dimensionalize(users.id) ;;
}

Input

• measure: A measure that you want to evaluate against a specific LoD context.
• dimension (repeatable): A dimension that you want to use as the LoD context.

Output

A new dimension

---

Sample Usages

The following example creates a dimension customer_lifetime_value, by dimensonalizing the measure total sales

dimension customer_lifetime_value {
  ...
  definition: @aql sum(orders.amount) | dimensionalize(users.id) ;;
} 

Total orders amount of this user equal to customer lifetime value field

We can also reuse this dimension just like other dimension to derive further analytics insight. Let's define average_customer_lifetime_value metric

metric average_customer_value {
  ...
  definition: @aql  average(users.customer_lifetime_value) ;;
}

Using the above metric to find Average Customer Lifetime Value per Continent