SQL Server：选择一列的计数，同时检查另一列中的不同值

Posted 2023-03-24

【中文标题】SQL Server：选择一列的计数，同时检查另一列中的不同值

【英文标题】：SQL Server : select count of one column, while checking for distinct value in another

【发布时间】：2017-05-29 17:55:45

【问题描述】：

我正在尝试获取某个列 (sch1.[key]) 的计数，同时还要确保另一列具有不同的值 (p.[ref])。有没有一种简单的方法可以使用 SQL Server 完成此任务？

我认为答案在于加入 [test] 和 [person] 表

测试表：记录列将匹配人员表 id

id     |    record    |   value
----------------------------------
01           901          abc
02           902          def
03           903          ghi

这是我现在所拥有的，但交叉应用并没有删除重复项

declare @dataset_orgs uniqueidentifier = (select top 1 [id] from [dataset] where ([name] = 'Organizations'))
select
  count(sch1.[key]) as [Total],
  sch1.[key] as [CEEB],
  sch1.[name] as [Institution],
  sch1dadd.[region] as [Region],
  (select [value] from dbo.getPromptTable(sch1dadd.[geomarket])) as [Geomarket],
  (select [value] from dbo.getFieldTopTable(sch1d.[id], 'staff_assign')) as [Staff]
from [test] t
CROSS APPLY
(
  select 
    DISTINCT p2.[id],
  from [person] p2
  where (p2.[id] = t.[record])
) p

inner join [lookup.test] lt on (lt.[id] = t.[type]) and (isnull(lt.[subtype], '') = isnull(t.[subtype], ''))
left outer join [school] sch1 on (sch1.[record] = p.[id]) and (sch1.[rank_overall] = 1)
left outer join [dataset.row] sch1d on (sch1d.[dataset] = @dataset_orgs) and (sch1d.[key] = sch1.[key])
left outer join [address] sch1dadd on (sch1dadd.[record] = sch1d.[id]) and (sch1dadd.[rank_overall] = 1)
where
(
  (t.[type] IN ('SATI', 'SATR'))
  and
  (convert(date, t.[date]) between isnull('1/1/2015', '1/1/1900') and isnull('12/31/2015', '12/31/9999'))
  and
  (isnull(convert(varchar(1), t.[confirmed]), '2') IN ('1'))
)
group by sch1.[key], sch1.[name], sch1dadd.[region], sch1dadd.[geomarket], sch1d.[id]
order by [Total] DESC`

【参考方案1】：

虽然你的查询不是很清楚，但据我了解你想要这样的东西

SELECT COUNT(*) AS CountAll, COUNT(DISTINCT p.ref) AS CountDistinct
FROM ...