MySQL query optimisation, SQL Query is taking too much time

Question

select * 
from `attendance_marks` 
where exists (select * 
              from `attendables` 
              where `attendance_marks`.`attendable_id` = `attendables`.`id` 
              and `attendable_type` = 'student' 
              and `attendable_id` = 258672 
              and `attendables`.`deleted_at` is null
              ) 
and (`marked_at` between '2022-09-01 00:00:00' and '2022-09-30 23:59:59')

this query is taking too much time approx 7-10 seconds. I am trying to optimize it but stuck at here.

Attendance_marks indexes

Attendables Indexes

Please help me optimize it a little bit.

For reference

number of rows in attendable = 80966

number of rows in attendance_marks = 1853696

Explain select

Answer 1

I think if we use JOINS instead of Sub-Query , then it will be more performant. Unfortunately, I don't have the exact data to be able to benchmark the performance.

select * 
from attendance_marks 
inner join attendables on attendables.id = attendance_marks.attendable_id
where attendable_type = 'student'
and attendable_id = 258672
and attendables.deleted_at is null
and (marked_at between '2022-09-01 00:00:00' and '2022-09-30 23:59:59')

I'm not sure if your business requirement allows changing the PK, and adding index. Incase it does then:

1. Add index to attendable_id .
2. I assume that attendables.id is PK. Incase not, add an index to it. Or preferably make it the PK.
3. In case attendable_type have a lot of different values. Then consider adding an index there too.
4. If possible don't have granularity till the seconds' field in marked_at , instead round to the nearest minute. In our case, we can round off 2022-09-30 23:59:59 to 2022-10-01 00:00:00 .

Answer 2

select  b.*
    from  `attendance_marks` AS am
    JOIN  `attendables` AS b  ON am.`attendable_id` = b.`id`
    WHERE  b.`attendable_type` = 'student'
      and  b.`attendable_id` = 258672
      and  b.`deleted_at` is null
      AND  am.`marked_at` >= '2022-09-01'
      AND  am.`marked_at` <  '2022-09-01 + INTERVAL 1 MONTH 

and have these

am:  INDEX(marked_at, attendable_id)
am:  INDEX(attendable_id, marked_at)
b:  INDEX(attendable_type, attendable_id, attendables)

Note that the datetime range works for any granularity.

(Be sure to check that I got the aliases for the correct tables.)

This formulation, with these indexes should allow the Optimizer to pick which table is more efficient to start with.