SQL必知必会--分组数据

创建分组：

利用group by分组  select count(*)from [Test].[dbo].[Products]group by [vend_id]

注：count是对每一个[vend_id]列进行计算，而不会整个表

       group by 有一个原则,就是 select 后面的所有列中,没有使用聚合函数的列,必须出现在 group by 后面，所以,select count(*),count([prod_id]),[vend_id],[prod_desc]
from [Test].[dbo].[Products] group by [vend_id]是错误的

过滤分组：

having 子句的作用是筛选满足条件的组。

select [vend_id],[prod_desc],count(*)from [Test].[dbo].[Products] group by [vend_id],[prod_desc] having count(*)>1

注：在分组之后过滤数据，条件中经常包含聚组函数，使用having 条件显示特定的组，也可以使用多个分组标准进行分组。

having和where的区别：

select [vend_id],[prod_desc],count(*)from [Test].[dbo].[Products] where [prod_price] > 30 group by [vend_id],[prod_desc] having count(*)>1

分组和排序：

区别：

排序：

select [vend_id],[prod_desc],count(*)from [Test].[dbo].[Products] where [prod_price] > 30  group by [vend_id],[prod_desc] order by[vend_id],[prod_desc]
select [vend_id],[prod_desc],count(*)from [Test].[dbo].[Products] where [prod_price] > 30 group by [vend_id],[prod_desc]

结果：

select子句顺序：