Types

Overview

There are 12 measure types that determine how values are aggregated. Choose the right type based on your analytical needs.

Counts the number of rows.

- name: count
  type: count

Counts unique values of a column.

- name: unique_users
  type: count_distinct
  sql: user_id

Approximate count distinct (faster for large datasets).

- name: approx_unique_users
  type: count_distinct_approx
  sql: user_id

Adds up numeric values.

- name: total_revenue
  type: sum
  sql: amount

Calculates the average of numeric values.

- name: average_order_value
  type: avg
  sql: amount

Returns the minimum value.

- name: first_order_date
  type: min
  sql: created_at

Returns the maximum value.

- name: last_order_date
  type: max
  sql: created_at

A calculated number (not aggregated from rows).

- name: conversion_rate
  type: number
  sql: "\{completed_orders\} / NULLIF(\{total_orders\}, 0)"

A calculated string value.

- name: status_label
  type: string
  sql: "'Active'"

A calculated time value.

- name: current_timestamp
  type: time
  sql: NOW()

A calculated boolean value.

- name: has_orders
  type: boolean
  sql: "\{count\} > 0"

Custom aggregate function (advanced).

- name: median_price
  type: number_agg
  sql: "PERCENTILE_CONT(0.5) WITHIN GROUP (ORDER BY \{CUBE\}.price)"