Question

我写了一个如下所示的查询：

    SELECT
id,
product_name,
CASE WHEN by_month = 'January 2018' THEN total_sales ELSE 0 END AS january_2018,
CASE WHEN by_month = 'February 2018' THEN total_sales ELSE 0 END AS february_2018,
CASE WHEN by_month = 'March 2018' THEN total_sales ELSE 0 END AS march_2018,
CASE WHEN by_month = 'April 2018' THEN total_sales ELSE 0 END AS april_2018
FROM (
SELECT
mnth,
id,
product_name,
CASE WHEN mnth = '2018-01-01' THEN 'January 2018' 
WHEN mnth = '2018-02-01' THEN 'February 2018' 
WHEN mnth = '2018-03-01' THEN 'March 2018' 
WHEN mnth = '2018-04-01' THEN 'April 2018' END AS by_month,
SUM(sales) total_sales
FROM (
SELECT
date_trunc('month', event_date) mnth,
id,
product_name,
SUM(sales) sales
FROM
table1
WHERE event_date BETWEEN '2017-01-01' AND '2018-04-30'
1,2,3)
WHERE sales > 0 
GROUP BY 
1,2,3,4)
GROUP BY 
1,2, by_month, total_sales

但这会产生一个非常零散的结果，看起来像这样

id   product_name january_2018 february_2018 march_2018 april 2018
123  abc               500           0           0          0
123  abc               0             700         0          0
123  abc               0             0         900          0
123  abc               0             0           0         100

我想要的是消除'0'结果（我知道这是我故意在ELSE语句中添加但我需要一个占位符。

我希望结果看起来像这样。

id   product_name january_2018 february_2018 march_2018 april 2018
123  abc               500          700         900         100

如何实现这一目标？它会出现在ELSE语句中还是其他子查询中？或其他什么？

Answer 1

使用聚合：

SELECT id, product_name,
       SUM(CASE WHEN by_month = 'January 2018' THEN total_sales ELSE 0 END) AS january_2018,
       SUM(CASE WHEN by_month = 'February 2018' THEN total_sales ELSE 0 END) AS february_2018,
       SUM(CASE WHEN by_month = 'March 2018' THEN total_sales ELSE 0 END) AS march_2018,
       SUM(CASE WHEN by_month = 'April 2018' THEN total_sales ELSE 0 END) AS april_2018
FROM table_name
GROUP BY id, product_name

Answer 2

对于更清晰的聚合，您可以将FILTER子句与聚合函数一起使用，如下所示：

SELECT
  id,
  product_name,
  SUM(total_sales) FILTER (WHERE by_month = 'January 2018') AS january_2018,
  SUM(total_sales) FILTER (WHERE by_month = 'February 2018') AS february_2018,
  SUM(total_sales) FILTER (WHERE by_month = 'March 2018') AS march_2018,
  SUM(total_sales) FILTER (WHERE by_month = 'April 2018') AS april_2018
FROM
  table_name
GROUP BY
  id,
  product_name

它是SQL：2003标准功能，但目前仅支持PostgreSQL。有关详细信息，请参阅The FILTER clause: Selective Aggregates

清理我正在尝试在SQL中创建的时间轴

2 个答案: