提取与转换日期/时间数据

在 PostgreSQL 中使用函数处理数据

Brian Piccolo

Sr. Director, Digital Strategy

提取与转换日期和时间数据

了解 EXTRACT()、DATE_PART() 和 DATE_TRUNC() 函数

交易时间戳精度不利于分析

2005-05-13 08:53:53

常需要提取时间戳的部分

2005 或 5 或 2 或 Friday

或转换/截断精度以标准化

2005-05-13 00:00:00

提取与转换日期/时间数据

EXTRACT( field FROM source )

SELECT EXTRACT(quarter FROM timestamp '2005-01-24 05:12:00') AS quarter;

DATE_PART('field', source)

SELECT DATE_PART('quarter', timestamp '2005-01-24 05:12:00') AS quarter;

+---------+
| quarter |
|---------|
| 1       |
+---------+

从时间戳提取子字段

DVD Rentals 的交易数据：payment 表

SELECT * FROM payment;

+--------------------------------------------------------------------------------+
| payment_id | customer_id | staff_id | rental_id | amount | payment_date        |
|------------|-------------|----------|-----------|--------|---------------------|
| 1          | 1           | 1        | 76        | 2.99   | 2005-05-25 11:30:37 |
| 2          | 1           | 1        | 573       | 0.99   | 2005-05-28 10:35:23 |
| 3          | 1           | 1        | 1185      | 5.99   | 2005-06-15 0:54:12  |
+--------------------------------------------------------------------------------+

从时间戳提取子字段

按年与季度汇总 payment 表数据

SELECT 
  EXTRACT(quarter FROM payment_date) AS quarter,
  EXTRACT(year FROM payment_date) AS year, 
  SUM(amount) AS total_payments 
FROM 
  payment 
GROUP BY 1, 2;

结果

+---------------------------------+
| quarter | year | total_payments |
|---------|------|----------------|
| 2       | 2005 | 14456.31       |
| 3       | 2005 | 52446.02       |
| 1       | 2006 | 514.18         |
+---------------------------------+

用 DATE_TRUNC() 截断时间戳

DATE_TRUNC() 用于截断时间戳或区间类型。

按"年"截断时间戳 '2005-05-21 15:30:30'

SELECT DATE_TRUNC('year', TIMESTAMP '2005-05-21 15:30:30');

结果: 2005-01-01 00:00:00

按"月"截断时间戳 '2005-05-21 15:30:30'

SELECT DATE_TRUNC('month', TIMESTAMP '2005-05-21 15:30:30');

结果: 2005-05-01 00:00:00

Passons à la pratique !

在 PostgreSQL 中使用函数处理数据