一个SQL查询及其相反的总和与总计数不匹配

Question

以下2个查询的总和为42081：

查询1：

SELECT COUNT(DISTINCT SIREN) AS new_column FROM CDEdelux
WHERE Email IS NOT NULL
AND Email <> ""
AND Email <> " "
AND Email <> ''
AND Email <> ' '
AND Email LIKE "%@%.%"
AND Email NOT LIKE "_@%.%"
AND Email NOT LIKE "%bpi%"
AND Email NOT LIKE "%BPI%"
AND Email NOT LIKE "%inconnu%"
AND Email NOT LIKE "%tempo%"
AND Email NOT LIKE "%attente%"
AND Email NOT LIKE "%xx%"
AND Email NOT LIKE "%nsp%"
AND Email NOT LIKE "%contact%"
AND Email NOT LIKE "%info%"
AND Email NOT LIKE "%recuperer%"

查询2（相反的查询）：

SELECT COUNT(DISTINCT SIREN) AS new_column FROM CDEdelux
WHERE Email IS NOT NULL
AND Email <> ""
AND Email <> " "
AND Email <> ''
AND Email <> ' '
AND (Email NOT LIKE "%@%.%"
OR Email LIKE "_@%.%"
OR Email LIKE "%bpi%"
OR Email LIKE "%BPI%"
OR Email LIKE "%inconnu%"
OR Email LIKE "%tempo%"
OR Email LIKE "%attente%"
OR Email LIKE "%xx%"
OR Email LIKE "%nsp%"
OR Email LIKE "%contact%"
OR Email LIKE "%info%"
OR Email LIKE "%recuperer%")

但是，由于它们应该是对立且互补的，因此应该给我39 206，这是该查询的结果：

查询3（总查询）：

SELECT COUNT(DISTINCT SIREN) AS new_column FROM CDEdelux
WHERE Email IS NOT NULL
AND Email <> ""
AND Email <> " "
AND Email <> ''
AND Email <> ' '

为什么第一个数字高于第二个？

Answer 1

好吧，很明显，给定的SIREN可能有多封符合任一条件的电子邮件。您可以通过以下方式查看重复的电子邮件：

select distinct siren, email
from CDEdelux d1
where exists (select 1 from CDEdelux d2 where d2.siren = d.siren and d2.email <> d.email);

如果您计数了不同的email，则这些数字应加起来。