Sql 如何组合第一个和第二个select语句

第一条select语句填充空值，第二条select语句删除重复项。如何从第一条SQL语句运行第二条SQL语句，或者是否有方法将它们组合起来

下面是SQL小提琴：

首先选择

/*
 Fill in null values with last record that is not null
*/
SELECT Id, FirstName, LastName, 

ISNULL(Address, 
(
SELECT TOP (1) Address    
 FROM test   
 WHERE (FirstName = MainTable.FirstName) AND (Address IS NOT NULL) AND (Address <> '')

ORDER BY Id DESC)

) AS Address

FROM test AS MainTable

 /*
Return a single record
*/
SELECT FirstName, LastName, MAX(Address) 
FROM test
Group By FirstName, LastName

预期结果：

| 4, john, doe, 9Street, SC | 
| 5, marry, doe, 78Street, DC | 

---

据我所知，您希望最新（按Id）地址和城市不为空，对吗？如果是：

SELECT max(t1.id) as id, t1.firstName, t1.lastName, adr.address, cty.City 
FROM test t1
cross apply (select top(1) address from test t2 
             where t1.FirstName = t2.FirstName and 
                   t1.LastName = t2.LastName and
                   t2.address is not null
             order by id desc) adr(Address)
cross apply (select top(1) city from test t2 
             where t1.FirstName = t2.FirstName and 
                   t1.LastName = t2.LastName and
                   t2.city is not null
             order by id desc) cty(City)
group by t1.firstName, t1.lastName, adr.address, cty.City;

---

如果您使用的是SQL Server 2012版或更高版本，则可以使用

last\u value（）

获取有序行（此处为

id

）分区（此处为

firstname

和

lastname

）中的最后一个值

不幸的是，它不允许忽略

NULL

s。但是，如果我们将感兴趣的列

为NULL的行排序在不为NULL的行之前，我们就可以解决这个问题。如果列
为空，我们可以首先按
大小写进行排序，如果列
为空，则返回一个较低的值；如果列
为空，则返回一个较高的值。第二列是
orderby
是
id

如果每个
firstname
和
lastname
只获取一行，那么简单的
DISTINCT
就可以了，因为
地址和
城市和
在共享
firstname
和
lastname
的任何行中都是相同的

SELECT DISTINCT
       t1.firstname,
       t1.lastname,
       last_value(t1.address) OVER (PARTITION BY t1.firstname,
                                                 t1.lastname
                                    ORDER BY CASE
                                            WHEN t1.address IS NULL THEN
                                              -1
                                            ELSE
                                              1
                                            END,
                                            t1.id
                                    RANGE BETWEEN UNBOUNDED PRECEDING
                                          AND UNBOUNDED FOLLOWING) address,
       last_value(t1.city) OVER (PARTITION BY t1.firstname,
                                              t1.lastname
                                 ORDER BY CASE
                                            WHEN t1.city IS NULL THEN
                                              -1
                                            ELSE
                                              1
                                          END,
                                          t1.id
                                 RANGE BETWEEN UNBOUNDED PRECEDING
                                               AND UNBOUNDED FOLLOWING) city
       FROM test t1;

如果您使用的是
SQLServer2012
或更高版本，
sticky bit
answer看起来是最好的答案。如果您使用的是较低版本，则可以检查以下答案

SELECT 
CAST((CASE
  WHEN CHARINDEX('$', Id) > 0
  THEN RIGHT(Id, CHARINDEX('$', REVERSE(Id)) - 1)
  ELSE Id
  END
) AS INT) AS Id,
Firstname, Lastname,
(CASE
  WHEN CHARINDEX('$', Address) > 0
  THEN RIGHT(Address, CHARINDEX('$', REVERSE(Address)) - 1)
  ELSE Address
  END
) AS Address,
(CASE
  WHEN CHARINDEX('$', City) > 0
  THEN RIGHT(City, CHARINDEX('$', REVERSE(City)) - 1)
  ELSE City
  END
) AS City
FROM (

SELECT A.Firstname, A.Lastname, 
STUFF((
  SELECT '$' + CAST(B.Id AS VARCHAR)
  FROM test B 
  WHERE B.Firstname = A.Firstname
  AND
  B.Lastname = A.Lastname
  For XML Path('')),1,1,'') AS Id,
STUFF((
  SELECT '$' + B.Address
  FROM test B 
  WHERE B.Firstname = A.Firstname
  AND
  B.Lastname = A.Lastname
  For XML Path('')),1,1,'') AS Address,
STUFF((
  SELECT '$' + B.City
  FROM test B 
  WHERE B.Firstname = A.Firstname
  AND
  B.Lastname = A.Lastname
  For XML Path('')),1,1,'') AS City
FROM test A  
GROUP BY A.Firstname, A.Lastname
) T

我认为标题是错误的，应该是“如何获取不同列的最新值”

还希望您的表应该包含
EmpID
或
USERID
，以便可以在其上创建索引

或者，如果可能，表格可以进一步标准化

Create table Table1(Empid int pk,FirstName,LastName)

Create table Table2(id int identity(1,1),EMPID FK,Adress,city)

查询这两个表将非常有效

SELECT DISTINCT t.Firstname
    ,t.Lastname
    ,t1.Address
    ,t2.City
FROM #test t
CROSS APPLY (
    SELECT TOP 1 t1.Address
    FROM #test t1
    WHERE t1.Address IS NOT NULL
        AND t.Firstname = t1.Firstname
        AND t.Lastname = t1.Lastname
    ORDER BY t1.id DESC
    ) t1
CROSS APPLY (
    SELECT TOP 1 t2.City
    FROM #test t2
    WHERE t2.City IS NOT NULL
        AND t.Firstname = t2.Firstname
        AND t.Lastname = t2.Lastname
    ORDER BY t2.id DESC
    ) t2

你期望的结果是什么？| 4，john，doe，9Street，SC | | 5，Mary，doe，78Street，DC |基本上这两个select语句填充空值，然后删除重复项。