窗口函数

使用 SQL 分析业务数据

Michel Semaan

Data Scientist

窗口函数——概览

窗口函数：对与当前行相关的一组行进行运算
示例
- 计算累计和
- 获取前一行或后一行的值

累计和

累计和（Running total）：对变量以往数值的累加

示例

x    x_rt
---  ----
1    1
2    3
3    6
4    11
5    16

注册量累计和——查询

WITH reg_dates AS (
  SELECT
    user_id,
    MIN(order_date) AS reg_date
  FROM orders
  GROUP BY user_id),

  registrations AS (
  SELECT
    DATE_TRUNC('month', reg_date) :: DATE AS delivr_month,
    COUNT(DISTINCT user_id) AS regs
  FROM reg_dates
  GROUP BY delivr_month)


SELECT
  delivr_month,
  regs,
  SUM(regs) OVER (ORDER BY delivr_month ASC) AS regs_rt
FROM registrations
ORDER BY delivr_month ASC LIMIT 3;

注册量累计和——结果

delivr_month  regs  regs_rt
------------  ----  -------
2018-06-01    123    123
2018-07-01    140    263
2018-08-01    157    420

滞后 MAU——查询

WITH maus AS (
  SELECT
    DATE_TRUNC('month', order_date) :: DATE AS delivr_month,
    COUNT(DISTINCT user_id) AS mau
  FROM orders
  GROUP BY delivr_month)


SELECT
  delivr_month,
  mau,
  COALESCE(
    LAG(mau) OVER (ORDER BY delivr_month ASC),
  1) AS last_mau
FROM maus
ORDER BY delivr_month ASC
LIMIT 3;

滞后 MAU——结果

delivr_month  mau  last_mau
------------  ---  --------
2018-06-01    123  1
2018-07-01    226  123
2018-08-01    337  226

窗口函数

使用 SQL 分析业务数据