count(distinct xxx) 和 group by 做去重隐藏的坑

不说废话，直接上代码。

以下查询是基于mysql自带的country,city,countrylanguage三个样例表做的：

select count(*) from country
union all 
SELECT count(distinct Code, Name, Continent, Region, SurfaceArea, IndepYear, Population, LifeExpectancy, GNP, GNPOld, LocalName, GovernmentForm, HeadOfState, Capital, Code2)
FROM world.country
union all 
select count(*) from 
(
SELECT Code, Name, Continent, Region, SurfaceArea, IndepYear, Population, LifeExpectancy, GNP, GNPOld, LocalName, GovernmentForm, HeadOfState, Capital, Code2
FROM world.country
group by Code, Name, Continent, Region, SurfaceArea, IndepYear, Population, LifeExpectancy, GNP, GNPOld, LocalName, GovernmentForm, HeadOfState, Capital, Code2
) a;

查询结果如下：

发现使用count(distinct xxx)后的结果居然和group by 去重的数据不一致；

不用怀疑，就是NULL引起的，以下为证：

SELECT distinct Code, Name, Continent, Region, SurfaceArea, IndepYear, Population, LifeExpectancy, GNP, GNPOld, LocalName, GovernmentForm, HeadOfState, Capital, Code2
FROM world.country

注意序号

SELECT distinct Code, Name, Continent, Region, SurfaceArea, IndepYear, Population, LifeExpectancy, GNP, GNPOld, LocalName, GovernmentForm, HeadOfState, Capital, Code2
FROM world.country
where 
Code is not null and 
Name is not null and 
Continent is not null and 
Region is not null and 
SurfaceArea is not null and 
IndepYear is not null and 
Population is not null and 
LifeExpectancy is not null and 
GNP is not null and 
GNPOld is not null and 
LocalName is not null and 
GovernmentForm is not null and 
HeadOfState is not null and 
Capital is not null and 
Code2 is not null 

 

 到这里大家应该明白了吧

使用count(distinct xxx) 时会自动过滤字段为NULL的值，而group by 不会，所以两个结果不一致。