[英]SQL: Filter out rows where column value occurs more than once
I have a MySQL table that looks like this: 我有一个如下所示的MySQL表:
id | label
----------------
1 "john"
1 "henry"
1 "sara"
2 "henry"
3 "tim"
So a given id
can have multiple label
s. 因此,给定的
id
可以具有多个label
。 I want to keep only the rows where the id
has a single label
. 我只想保留
id
具有单个label
。 So the correct output for the above table would be: 因此,上表的正确输出为:
id | label
----------------
2 "henry"
3 "tim"
I was thinking I should group by id
and find the count of labels for each id
. 我当时想我应该按
id
分组并找到每个id
的标签数。 Then I'd take only rows with a count of 1. 然后,我只需要计数为1的行。
WITH temp as
(SELECT id
FROM original_table
GROUP BY id
HAVING COUNT(id) > 5)
SELECT *
FROM original_table ot
WHERE ot.id in temp.id
Does that look close? 看起来很近吗?
Thanks! 谢谢!
You could just use a join to only include ID's that occur once in a sub-query: 您可以只使用联接仅包含在子查询中出现一次的ID:
SELECT id,
label
FROM original_table ot
INNER JOIN (
SELECT id
FROM original_table
GROUP BY id
HAVING COUNT(*) = 1
) a ON a.id = ot.id;
Or you could use an IN
clause: 或者您可以使用
IN
子句:
SELECT id,
label
FROM original_table
WHERE id IN (SELECT id
FROM original_table
GROUP BY id
HAVING COUNT(*) = 1
);
I think aggregation is the simplest method: 我认为聚合是最简单的方法:
select id, min(label) as label
from original_table t
group by id
having count(*) = 1;
You could try this: 您可以尝试以下方法:
SELECT t.id, t.label
FROM tbl AS t
JOIN (SELECT id FROM tbl GROUP BY id HAVING count(label) = 1) AS t1
ON t.id = t1.id;
Assuming that the pair id
and label
is unique, you can use NOT EXISTS
and a correlated subquery. 假设对
id
和label
是唯一的,则可以使用NOT EXISTS
和相关的子查询。
SELECT t1.id,
t1.label
FROM original_table t1
WHERE NOT EXISTS (SELECT *
FROM original_table t2
WHERE t2.id = t1.id
AND t2.label <> t1.label);
Yes your approach is correct and you may have to change your having count condition and while referring to CTE you may have to change your syntax little , but you can do it without CTE as well in the same line with exists condition. 是的,您的方法是正确的,您可能必须更改您的计数条件,而在引用CTE时,您可能需要稍微更改语法,但是在存在条件的同一行中也可以不使用CTE来执行此操作。
Create table temp (ID int , Label varchar(10));
insert into temp values
(1 , "john" ),
(1 , "henry" ) ,
(1 , "sara" ) ,
( 2 , "henry" ) ,
(3 , "tim" ) ;
select t.ID , t.Label from temp t
where exists (
select ID, count(1) Dups from temp t1 where t1.ID = t.ID group by ID having count(1)
= 1)
Output: 输出:
ID, Label
2, henry
3, tim
声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.