假设我有一个电子邮件功能,我需要选择电子邮件的收件人。
users
表格包含以下列:id
,full_name
emails
表包含以下列:id
,subject
,body
,sender_id
,recipient_id
伪代码是
SELECT
e.id AS email_id,
e.subject AS email_subject,
u_sender.full_name AS sender_name,
u_recipient.full_name AS recipient_name
FROM emails e
LEFT JOIN users u_sender
ON e.sender_id = u_sender.id
LEFT JOIN users u_recipient
ON e.recipient_id = u_recipient.id
这将返回如下内容:
| email_id | email_subject | sender_name | recipient_name |
| 1
|你好
|约翰
|史蒂夫|
| 1
|你好
|约翰
|马克|
| 1
|你好
|约翰
|彼得|
相反,有没有办法得到类似的东西:
| email_id | email_subject | sender_name | recipient_name |
| 1
|你好
|约翰
| [史蒂夫,彼得,马克] |
请注意,它们具有相同的email_id,因为我的应用程序可以将电子邮件发送给多个收件人。
我找到a similar question,但解决方案是apache-exclusive。
我还发现了一个使用XML PATH here的响应,但它适用于SQL Server 2005.我正在使用postgres。
提前致谢。
答案 0 :(得分:2)
PostgreSQL 9提供string_agg
聚合函数,将多个值展平为单个字符串。在您的示例中,它将是string_agg(u_recipient.full_name, ',')
以及合适的group by
:
select
e.id email_id,
e.subject email_subject,
u_sender.full_name sender_name,
string_agg(u_recipient.full_name, ',') recipient_name
from emails e
left join users u_sender
on e.sender_id = u_sender.id
left join users u_recipient
on e.recipient_id = u_recipient.id
group by email_id, email_subject, sender_name
答案 1 :(得分:0)
将GROUP BY子句与email_id,email_subject,sender_name一起使用,并调用STRING_AGG(recipient_name,',')
你会有这样的事情:
SELECT email_id, email_subject, sender_name, string_agg(recipient_name, ',') FROM test GROUP BY email_id, email_subject, sender_name
使用您的查询:
WITH X AS
(
SELECT
e.id AS email_id,
e.subject AS email_subject,
u_sender.full_name AS sender_name,
u_recipient.full_name AS recipient_name
FROM emails e
LEFT JOIN users u_sender
ON e.sender_id = u_sender.id
LEFT JOIN users u_recipient
ON e.recipient_id = u_recipient.id
)
SELECT email_id, email_subject, sender_name, string_agg(recipient_name, ',') FROM X GROUP BY email_id, email_subject, sender_name;