sql - Combine two resulting rows in a query

Question

I have the follow query:

select user_id from user_tags
where user_id = 3086533

which outputs:

Essentially, I want to be able to take this query, and change it to only include the user_id if both the name "master_sbx" and "phonePicked"appear.

So instead of writing a query of:

select user_id from user_tags
where user_id = 3086533

with an output of:

I would like to add the clause of only including user_id of only ids that have both "master_sbx" and "phonePicked" applied, so the output would just return

Hoping this makes sense. I'm not sure its actually possible in PostGreSQL but would love any feedback! It is piece of a larger query I am going to write.

Answer 1

如果我理解正确，您想测试 auser_id是否（至少）出现在name = 'master_sbx'与name = 'phonePicked'.

有很多方法可以做到这一点。最佳选择取决于不在您问题中的信息。一种方法是EXISTS：

SELECT DISTINCT user_id
FROM   user_tags
WHERE  user_id = 3086533
AND    name = 'master_sbx'
AND    EXISTS (
   SELECT 1 FROM user_tags
   WHERE  user_id = 3086533
   AND    name = 'phonePicked'
   )

另一种自加入方式：

SELECT DISTINCT u1.user_id
FROM   user_tags u1
JOIN   user_tags u2 USING (user_id)
WHERE  u1.user_id = 3086533
AND    u1.name = 'master_sbx'
AND    u2.name = 'phonePicked';

DISTINCT如果(user_id, name)是唯一的，则是多余的。

这可以看作是关系划分的特例。我们在这个相关答案中组装了一整套查询技术：
How to filter SQL results in a has-many-through relationship