sql Continuous login problem

Today we have a video , Requirements are calculated in a library table , Inquire about 2022 year 1 Users who log in for three consecutive days in the month and the duration of continuous login .

First ,t_login It records the time information of each login .

Log in multiple times a day , Just record one ;

( Using functions invalidates the index )

select distinct uid,date(login_time) ymd
from t_login
where login_time between '2022-01-01 00:00:00' and timestamp '2022-01-31 23:59:59';

Rough way , Self associate the above table three times , shortcoming , Expandability of .

select t1.uid,t1.ymd,t2.ymd,t3.ymd 
from
(select distinct uid,date(login_time) ymd
from t_login
where login_time between '2022-01-01 00:00:00' and timestamp '2022-01-31 23:59:59') t1
join
(select distinct uid,date(login_time) ymd
from t_login
where login_time between '2022-01-01 00:00:00' and timestamp '2022-01-31 23:59:59') t2
on
(t1.uid=t2.uid and datediff(t2.ymd,t1.tmd)=1)
join
(select distinct uid,date(login_time) ymd
from t_login
where login_time between '2022-01-01 00:00:00' and timestamp '2022-01-31 23:59:59') t3
on
(t2.uid=t3.uid and datediff(t3.ymd,t2.tmd)=1);

Mode two

Use window functions , Partition restrictions , obtain

with t1 as (select distinct uid,date(login_time) ymd
            from t_login
            where login_time between timestamp '2022-01-01 00:00:00' 
            				and timestamp '2022-01-31 23:59:59')           				
select uid,ymd,
row_number() over (partition by uid order by ymd) num
from t1;          

As long as it's not continuous , There must be a result jump in interpolation .

with t1 as (select distinct uid,date(login_time) ymd
            from t_login
            where login_time between timestamp '2022-01-01 00:00:00' 
            				and timestamp '2022-01-31 23:59:59')  
t2 as(select uid,ymd,
date_sub(ymd,interval row_number() over (partition by uid order by ymd) day) sub_date
from t1)

select uid,min(ymd),max(ymd),count(*)
from t2
group by uid,sub_date
having count(*)>=3;

Fucking regret , It's a waste of time . Alas, I have no time to sigh .