连接两个子查询或查询使用子查询的SQL Server 2008(Joining two subque

怪物编辑：查询现在将运行，但将返回不正确的答案。 增加了一个粗略的架构。 PatientID不在tblPatientVisits表的主键，因为相同的patientID可以存在多于一次。

每个县名称上市，但每个计数（s.countyName）和计数（t.countyname）为1

PatientVisits   
PatientID   int
PatientState    varchar(2)
patientCounty   varchar(3)
visitNumber int - PK

tblStateCounties    
CPK stateCode   varchar(2)
CPK countyCode  varchar(3)
countyName  varchar(25)


SELECT t.countyName,
    count(t.countyName) as reAdmits ,count(s.countyName) as totalVisits
FROM (
    SELECT countyName,count(countyName) AS readmitCounts
    FROM (
        SELECT tblPatient.patientID
            ,tblStateCounties.countyName
        FROM tblPatient
        INNER JOIN tblPatientVisits
            ON tblPatient.patientID = tblPatientVisits.patientID
        INNER JOIN tblStateCounties
            ON tblPatientVisits.patientState = tblStateCounties.stateCode
                AND tblPatientVisits.patientCounty = tblStateCounties.countyCode
        GROUP BY tblPatient.patientID
            ,tblStateCounties.stateCode
            ,tblStateCounties.countyName
        HAVING (COUNT(tblPatient.patientID) > 1)
            AND (tblStateCounties.stateCode = '21')
        ) t
    GROUP BY countyname
    ) t
INNER JOIN (
    SELECT countyName
    FROM (
        SELECT tblStateCounties.countyName
            ,COUNT(tblStateCounties.countyName) AS counts
        FROM tblPatient
        INNER JOIN tblPatientVisits
            ON tblPatient.patientID = tblPatientVisits.patientID
        INNER JOIN tblStateCounties
            ON tblPatientVisits.patientState = tblStateCounties.stateCode
                AND tblPatientVisits.patientCounty = tblStateCounties.countyCode
        WHERE (tblStateCounties.stateCode = '21')
        GROUP BY tblStateCounties.countyName
        ) z
    ) s
    ON s.countyName = t.countyName
    group by s.countyname, t.countyname

编辑：我现在运行一个查询，但它返回

Answer 1:

很难说没有样本数据和预期的结果，但也许这是你所追求的？

;WITH x AS 
(
  SELECT c.CountyName, v.patientCounty, v.patientState, p.patientID
    FROM dbo.tblPatient AS p
    INNER JOIN dbo.tblPatientVisits AS v
    ON p.patientID = v.patientID 
    INNER JOIN dbo.tblStateCounties AS c 
    ON v.patientState = c.stateCode 
    AND v.patientCounty = c.countyCode
  WHERE c.stateCode = '21'
),
y AS (SELECT CountyName, c = COUNT(*) FROM x GROUP BY CountyName),
z AS (SELECT CountyName, c = COUNT(PatientID) FROM x 
  GROUP BY CountyName, patientState, PatientID HAVING COUNT(*)>1)
SELECT y.countyName, reAdmits = MAX(COALESCE(z.c, 0)), totalVisits = MAX(y.c)
FROM y LEFT OUTER JOIN z
ON y.CountyName = z.CountyName
GROUP BY y.CountyName;

Answer 2:

似乎有这个查询的几个问题。 我假定为“S”的子查询意图之前被替换为在查询的“s”。 你熟悉“与”语法？ 这将是接近你是如何表达它。

在任何情况下，你的第一个子查询缺少“）” HAVING子句和GROUP BY之间。 此外， “S” 子查询有GROUP BY，但 “SELECT *”。

以下可能是你想表达什么：

select t.countyName, count(t.countyName), s.countyName, count(s.countyName)
from (select countyName, count(countyName) as readmitCounts
      from (SELECT tblPatient.patientID, tblStateCounties.countyName
            FROM tblPatient INNER JOIN
                 tblPatientVisits
                 ON tblPatient.patientID = tblPatientVisits.patientID INNER JOIN
                 tblStateCounties
                 ON tblPatientVisits.patientState = tblStateCounties.stateCode AND
                    tblPatientVisits.patientCounty = tblStateCounties.countyCode
            GROUP BY tblPatient.patientID, tblStateCounties.stateCode, tblStateCounties.countyName
            HAVING (COUNT(tblPatient.patientID) > 1) AND (tblStateCounties.stateCode = '21')
           ) t
      group by countyname
     ) t inner join
     (select tblStateCounties.countyName
      from (SELECT tblStateCounties.countyName, COUNT(tblStateCounties.countyName) AS counts
            FROM tblPatient INNER JOIN
                 tblPatientVisits
                 ON tblPatient.patientID = tblPatientVisits.patientID INNER JOIN
                 tblStateCounties
                 ON tblPatientVisits.patientState = tblStateCounties.stateCode AND 
                    tblPatientVisits.patientCounty = tblStateCounties.countyCode
            WHERE (tblStateCounties.stateCode = '21')
       GROUP BY tblStateCounties.countyName
     )
     on s.countyName = t.countyName

Answer 3:

我觉得你过于复杂的问题，并认为这可以在不糖化2个查询一起做到：

SELECT  t.countyName, 
        SUM(Admissions) AS TotalAdmissions,
        COUNT(*) AS TotalPatients,
        COUNT(CASE WHEN Admissions > 1 THEN PatientID END) AS TotalPatientsReadmitted
        SUM(Admissions - 1) AS TotalReadmissions
FROM    (   SELECT  tblPatient.PatientID,
                    tblStateCounties.countyName,
                    COUNT(*) AS Admissions
            FROM    tblPatient 
                    INNER JOIN tblPatientVisits 
                        ON tblPatient.patientID = tblPatientVisits.patientID 
                    INNER JOIN tblStateCounties 
                        ON tblPatientVisits.patientState = tblStateCounties.stateCode 
                        AND tblPatientVisits.patientCounty = tblStateCounties.countyCode
            WHERE   tblStateCounties.stateCode = '21'
            GROUP BY tblPatient.PatientID, tblStateCounties.countyName
        ) AS t
GROUP BY CountyName

我刚拍摄的FROM和JOINS让我觉得什么是你所需要的（多一点）的数据。 我希望能够给列明智足够的名字为它是什么显而易见的每一个应该显示。

文章来源: Joining two subqueries or a query with a subquery SQL Server 2008