Question

我遇到了一个问题，我试图返回的列中只有1个值匹配的行。例如：

CREATE TABLE TEST
(
    ID INT PRIMARY KEY,
    FIRSTNAME VARCHAR(30) NOT NULL,
    LASTNAME VARCHAR(30) NOT NULL,
    EMAIL VARCHAR(50),
    PHONENUMBER VARCHAR(50)
    STAMP DATETIME
)

INSERT INTO TEST VALUES (1, 'Tom', 'Lew', 'tom@hotmail.com', '11111111111', '2019-08-23 15:12:06.807')
INSERT INTO TEST VALUES (2, 'Tom', 'Lew', 'tom@hotmail.com', '22222222222', '2019-08-27 15:12:06.807')
INSERT INTO TEST VALUES (3, 'Jack', 'Stan', 'jstan@hotmail.com', '3333333333', '2019-08-03 15:12:06.807')
INSERT INTO TEST VALUES (4, 'John', 'Doe', 'jdoe@hotmail.com',  '44444444444', '2019-08-13 15:12:06.807')
INSERT INTO TEST VALUES (5, 'Peter', 'Griffin', 'pgriffin@hotmail.com', '55555555555', '2019-07-23 15:12:06.807')
INSERT INTO TEST VALUES (6, 'Homer', 'Simpson', 'hsimpson@hotmail.com', '66666666666', 2019-08-23 15:12:06.807')

SELECT FirstName, LastName, Email, PhoneNumber
FROM TEST GROUP BY FirstName, LastName, Email, PhoneNumber
HAVING COUNT (FirstName) <= 1

因此，应该假定结果返回除“ Tom”之外的所有内容，因为“ Tom”在“名字”列中具有多个值匹配。但是，它正在返回所有内容。

如果我执行SELECT FirstName，LastName，Email和GROUP BY FirstName，LastName，Email，它会正确返回。

Answer 1

您可以使用not exists。如果您只关心名字：

select t.*
from test t
where not exists (select 1
                  from t t2
                  where t2.firstname = t.firstname and
                        t2.id <> t.id
                 );

Answer 2

您选择查询的问题是两个汤姆没有相同的电话号码。如果您要在GROUP BY子句中包含电话号码，它将为您提供两个唯一的两行。您需要确定GROUP BY的唯一组合列。

如果您只关心名字，您的查询将是这样。

SELECT FirstName, LastName, Email, PhoneNumber
FROM TEST GROUP BY FirstName
HAVING COUNT (FirstName) <= 1

如果您将 LastName 和 Email 放入GROUP BY中，则：

SELECT FirstName, LastName, Email, PhoneNumber
FROM TEST GROUP BY FirstName, LastName, Email
HAVING COUNT (FirstName) <= 1

Answer 3

由于您检查唯一的FirstName，因此您只应GROUP BY FirstName并使用MAX()或MIN()在其他列上进行汇总（这没有任何区别，因为条件HAVING子句中，请确保这些列中只有1个值存在）：

SELECT FirstName, 
  MAX(LastName) LastName, MAX(Email) Email, MAX(PhoneNumber) PhoneNumber
FROM TEST 
GROUP BY FirstName
HAVING COUNT(*) = 1

请参见demo。
结果：

| FirstName | LastName | Email                | PhoneNumber |
| --------- | -------- | -------------------- | ----------- |
| Homer     | Simpson  | hsimpson@hotmail.com | 66666666666 |
| Jack      | Stan     | jstan@hotmail.com    | 3333333333  |
| John      | Doe      | jdoe@hotmail.com     | 44444444444 |
| Peter     | Griffin  | pgriffin@hotmail.com | 55555555555 |

Answer 4

我将使用ROW_NUMBER（）排除重复的记录：

SELECT * FROM(
              SELECT firstname, lastname, email, phonenumber,
              ROW_NUMBER() OVER(PARTITION BY firstname ORDER BY firstname) AS rn
              FROM test
            )
 WHERE rn=1;

希望这会有所帮助。

SQL-如果列的值不相同，则仅返回行

4 个答案: