印象中这两个 SQL 查询会给出相同的输出，但结果却截然不同

Question

我正在使用 pandasql。第一个按预期转换值，但第二个 return 甚至不应该存在的东西。但是，我希望它们 return 具有相同的值。在我看来，唯一的区别是，在第一个中， grouping/sum 出现在子查询中，而在第二个中它出现在子查询之外。我错过了什么？感谢您的帮助！（输出在底部）

第一个查询（正确的）

SELECT a.'Name', a.Q1, b.Q2, (a.Q1 + b.Q2) AS Total
FROM
    (SELECT c.'Name', SUM(c.'Paid Amount') AS Q1
     FROM some_data AS c
     WHERE c.'Quarter' = 'Q1'
     GROUP BY c.'Name') AS a
JOIN
    (SELECT d.'Name', SUM(d.'Paid Amount') AS Q2
     FROM some_data AS d
     WHERE d.'Quarter' = 'Q2'
     GROUP BY d.'Name') AS b
ON a.'Name' = b.'Name'
ORDER BY Total DESC
LIMIT 5;

第二次查询（不好的）

SELECT a.'Name' as Label, SUM(a.'Paid Amount') AS Q1, SUM(b.'Paid Amount') AS Q2, (SUM(a.'Paid Amount') + SUM(b.'Paid Amount')) as Total
FROM 
    (SELECT c.'Name', c.'Paid Amount'
     FROM some_data AS c
     WHERE c.'Quarter' = 'Q1') AS a
JOIN
    (SELECT c.'Name', c.'Paid Amount'
     FROM some_data AS c
     WHERE c.'Quarter' = 'Q2') AS b
ON a.'Name' = b.'Name'
GROUP BY Label
ORDER BY Total DESC
LIMIT 5;

我把一些随机数据放在一起来证明这个问题。

第一个查询的输出（预期）

第二次查询的输出（有问题）

Answer 1

这就是我所说的一厢情愿的编码。

我希望您意识到在加入之前进行聚合会产生正确的答案。

问题是 JOIN 既可以乘以行数也可以删除行。在您的情况下，问题是一个或两个表都有 name 的多行，这会乘以行数。 SUM() 只是将 JOIN.

产生的所有值相加

注意：条件聚合是一种更简单的查询编写方式：

SELECT c.Name,
       SUM(CASE WHEN c.Quarter = 'Q1' THEN c.PaidAmount END) AS Q1
       SUM(CASE WHEN c.Quarter = 'Q2' THEN c.PaidAmount END) AS Q2
FROM some_data AS c
WHERE c.Quarter IN ('Q1', 'Q2')
GROUP BY c.Name

印象中这两个 SQL 查询会给出相同的输出，但结果却截然不同

Under the impression that these two SQL queries would give the same output, yet they have wildly different results

sql

grouping

pandas

pandasql