MYSQL select to get Consecutive Day Count user wise where the value is lesser than previous day value - select

Question

MySql V 8.0
Question: How to write MySQL select to get Consecutive Day Count where the weight value is lesser than the previous day weight value user wise and break when no longer consecutive or weight value is same or greater than the previous day weight value of the same user.
create table userData (recordDate ,userName varchar(100), weight FLOAT);
insert into userData (recordDate, userName, weight)
values
('2020/8/1','Chris', 78),
('2021/8/2','Chris', 77),
('2021/8/3','Chris', 76),
('2021/8/1','Aamir', 78),
('2021/8/2','Aamir', 77),
('2021/8/1','Alex', 78),
('2021/8/2','Alex', 77),
('2021/8/3','Alex', 76),
('2021/8/5','Chris', 78),
('2021/8/6','Chris', 77),
('2021/8/7','Chris', 76),
('2021/8/8','Chris', 75),
('2021/8/8','Aamir', 78),
('2021/8/8','Alex', 78),
('2021/8/9','John', 78),
('2021/8/1','Ali', 78),
('2021/8/10','Chris', 78);
The expected output is
| userName | streakDays | startingDate | endingDate |
| -------- | ---------- | ------------ | ---------- |
| Alex | 3 | 2021-08-01 | 2021-08-03 |
| Chris | 3 | 2021-08-06 | 2021-08-08 |
| Aamir | 2 | 2021-08-01 | 2021-08-02 |
| Ali | 1 | 2021-08-01 | 2021-08-01 |
| John | 1 | 2021-08-09 | 2021-08-09 |
Any help would be appreciated.

Answer 1

According To Your data inserted in the table , This Select Query Works Fine
select userName as un ,
count((select recordDate WHERE userName = un)) as strekdays,
(select recordDate FROM userdata WHERE userName = un limit 1) as startdate ,
(select recordDate FROM userdata WHERE userName = un order by recordDate DESC limit 1) as enddate
from userdata
group by userName
And It Gives Output Like
userName
streakDays
startingDate
endingDate
Aamir
3
2021-08-01
2021-08-08
Alex
4
2021-08-01
2021-08-08
Ali
1
2021-08-01
2021-08-01
Chris
8
2021-08-01
2021-08-10
John
1
2021-08-09
2021-08-09
Let me know If this Works FOr You or not !

Answer 2

Problem resolved with the following query:
select
streakBreakersRemoved.userName,
streakBreakersRemoved.streakDays,
streakBreakersRemoved.startingDate,
streakBreakersRemoved.endingDate
from
(
select
userName,
count(*) as streakDays,
min(recordDate) as startingDate,
max(recordDate) as endingDate,
row_number() over (partition by userName
order by
count(*) desc) as seqNum
from
(
select
initailRecords.*,
row_number() over (partition by userName
order by
recordDate) as initialSeqNum
from
(
select
userData.*,
lag(weight) over (partition by userName
order by
recordDate) as previousWeight
from
userData
)
initailRecords
where
if(previousWeight is null || previousWeight > weight, 1, 0) = 1
)
recordsWithSeqNum
group by
userName,
to_days(recordDate) - initialSeqNum
)
streakBreakersRemoved
where
seqNum = 1
order by
streakDays desc;
Would appreciate if anyone would like to optimize the above query.