I have this table:
+-----------+-----------+---------------------+
| id | member_id | date |
+-----------+-----------+---------------------+
| 1 | 2 | 2020-07-27 21:53:46 |
| 2 | 1 | 2020-07-27 22:03:58 |
| 3 | 1 | 2020-07-27 22:09:16 |
| 4 | 1 | 2020-07-27 22:11:33 |
| 5 | 2 | 2020-07-27 22:12:21 |
-----------------------------------------------
And I would like to GROUP BY something like this:
+-----------+-----------+---------------------+
| id | member_id | date |
+-----------+-----------+---------------------+
| 1 | 2 | 2020-07-27 21:53:46 |
| 2 | 1 | 2020-07-27 22:03:58 |
| 5 | 2 | 2020-07-27 22:12:21 |
-----------------------------------------------
But when I put GROUP BY member_id it just return me two rows (1, 2)
I understand this as a gaps-and-island problem, where you want to display the rows whose member_id is different than the "previous" row.
Here is an approach using window functions (available in MySQL 8.0): lag() gives the member_id on the previous row, that we can then compare to the value on the current row:
select id, member_id, date
from (
select t.*, lag(member_id) over(order by date) lag_member_id
from mytable t
) t
where not lag_member_id <=> member_id
order by date
id | member_id | date -: | --------: | :------------------ 1 | 2 | 2020-07-27 21:53:46 2 | 1 | 2020-07-27 22:03:58 5 | 2 | 2020-07-27 22:12:21