Question

我在PostgreSQL引擎驱动的数据库中有一个表answers，其中一个数字与一个问题相关联。我需要计算有多少数字，以及有多少数字低于6，按问题分组。

我想要的是：

SELECT question, count(*) AS Qanswers, count_below_6(*) AS Qanswers_below_6
FROM answers
GROUP BY question;

      question         | Qanswers | Qanswers_below_6
-----------------------+----------+------------------
How do you feel?       |  1234    |      53
Was clear the webinar? |  8444    |      20
How much that hurt?    |  3666    |     142

目前我正在做

SELECT question, count(*) AS Qanswers
FROM answers
GROUP BY question;

然后

SELECT question, count(*) AS Qanswers
FROM answers
WHERE value < 6
GROUP BY question;

之后我手动合并两个结果我可以根据需要制作一个能给我结果的单句吗？

Answer 1

SELECT question
      ,count(*) AS answers
      ,count(value < 6 OR NULL) AS under_6
FROM   answers
GROUP  BY question;

诀窍是count(expression)仅计算非空值（而count(*)计算所有行）。这里详细解释了为什么我们在正确的地方得到NULL：
Compute percents from SUM() in the same SELECT sql query

有许多等效技术可以实现这一目标。与基准的详细比较：
For absolute performance, is SUM faster or COUNT?

计算总数和部分数

1 个答案: