MySQL - GROUP multiple rows while retaining differences

问题

I have a series of complicated MySQL select queries / joins that produce results similar to this:

|----|---------|----------|----------|---------------|---------------|
| id | company | city     | province | manager_email | staff_email   |
|----|---------|----------|----------|---------------|---------------|
| 1  | aaa     | toronto  | ON       | john@aaa.com  |               |
| 1  | aaa     | toronto  | ON       |               | smith@aaa.com |
| 2  | bbb     | sudbury  | ON       | john@bbb.com  |               |
| 3  | ccc     | hamilton | ON       | john@ccc.com  |               |
| 3  | ccc     | hamilton | ON       |               | smith@ccc.com |
|----|---------|----------|----------|---------------|---------------|

• most "companies" have two rows, which are identical other than the emails
• manager_email and staff_email will never appear in the same row
• there are some cases where a "company" will only have one row

Is there a GROUP BY or a similar statement that I can use to group all of these duplicate rows, while retaining both emails? e.g.

|----|---------|----------|----------|---------------|---------------|
| id | company | city     | province | manager_email | staff_email   |
|----|---------|----------|----------|---------------|---------------|
| 1  | aaa     | toronto  | ON       | john@aaa.com  | smith@aaa.com |
| 2  | bbb     | sudbury  | ON       | john@bbb.com  |               |
| 3  | ccc     | hamilton | ON       | john@ccc.com  | smith@ccc.com |
|----|---------|----------|----------|---------------|---------------|

I'm willing to share more details if needed, but at this point I think it'll just add confusion.

回答1:

You can take advantage of the fact that most aggregate functions ignore null:

select 
    id, 
    company, 
    city, 
    province, 
    max(manager_email) manager_email, 
    max(staff_email) staff_email
from mytable
group by id, company, city, province

来源：https://stackoverflow.com/questions/58957478/mysql-group-multiple-rows-while-retaining-differences