SQL select students logins - sql

Question

I have a table STUDENT_LAST_LOGIN, which contains data about students last logins.
ID STUDENT_ID DATE TIME
1 A 2020-02-01 12:00 15 MIN
2 B 2020-02-02 12:00 45 MIN
3 C 2020-02-03 12:00 25 MIN
In addition there is STUDENT_LOGIN table, which contains data about students all logins.
ID STUDENT_ID DATE TIME
1 A 2020-02-01 12:00 15 MIN
4 A 2020-01-01 14:00 33 MIN
2 B 2020-02-02 12:00 45 MIN
5 B 2020-01-02 13:30 47 MIN
10 B 2020-01-03 13:30 27 MIN
6 B 2020-01-02 10:00 44 MIN
3 C 2020-02-03 12:00 25 MIN
7 C 2020-01-03 10:00 12 MIN
8 C 2020-01-03 18:00 56 MIN
9 C 2020-01-04 12:00 88 MIN
As a result I need to get something like this:
STUDENT_ID LAST_LOGIN LAST_LOGIN_ONE_MONTH_AGO TIME TIME_ONE_MONTH_AGO
A 2020-02-01 12:00 2020-01-01 14:00 15 min 33 min
B 2020-02-02 12:00 2020-01-02 13:30 15 min 47 min
C 2020-02-03 12:00 2020-01-03 18:00 25 min 56 min
Can you help me write this?

Answer 1

SELECT LAST_LOGIN, LAST_LOGIN_ONE_MONTH_AGO, S_L.TIME, S_L.TIME_ONE_MONTH_AGO
FROM STUDENT_LAST_LOGIN S_L_L
INNER JOIN STUDENT_LOGIN S_L on S_L_L.id = S_L.id
where S_L_L.date < DATEADD(month, -1, GETDATE())
you need to write your query something like this.

Answer 2

You need to use the windows function as follows:
SELECT * FROM
(SELECT SLL.STUDENT_ID,
SLL.DATE LAST_LOGIN,
SL.DATE LAST_LOGIN_ONE_MONTHE_AGO,
SLL.TIME,
SL.TIME TIME_ONE_MONTH_AGO,
ROW_NUMBER() OVER (PARTITION BY SLL.STUDENT_ID ORDER BY SL.DATE DESC NULLS LAST) AS RN
FROM STUDENT_LAST_LOGIN SLL LEFT JOIN STUDENT_LOGIN SL
ON SL.STUDENT_ID = SLL.STUDENT_ID
AND TRUNC(SL.DATE) = ADD_MONTHS(TRUNC(SLL.DATE),-1)
)
WHERE RN = 1

Answer 3

I can only speculate that you want the most recent login and then the most recent login from the calendar month before that. I would suggest conditional aggregation:
select sll.student_id,
max(case when month_seqnum = 1 then last_login end),
max(case when month_seqnum = 2 then last_login end),
max(case when month_seqnum = 1 then time end),
max(case when month_seqnum = 2 then time end)
from (select sll.*,
row_number() over (partition by student_id, to_char(date, 'YYYY-MM')
order by date desc
) as seqnum,
dense_rank() over (partition by student_id order by to_char(date, 'YYYY-MM')) as month_seqnum
from student_last_login sll
) sll
where month_seqnum in (1, 2) and seqnum = 1
group by student_id;
I think this returns the values that you specify.