Question

I’m working on an internal signup form for classes. I am working on a simple function that will check if the user has selected classes with conflicting schedules ( E.g. A class on from 1-2 p.m. as well as a class from 1-3 p.m. on the same day.)

Assuming that I’m using mysqli and fetching the classes the user has selected from a MySQL table with start and end datetime fields, what is the most efficient and/or effective way to compare the multiple classes in order to see if they conflict?

EDIT: Here’s a sample table:

====================================================
|  id  |        start        |         end         | 
====================================================
|  1   | 2012-10-01 08:00:00 | 2012-10-01 08:00:00 |
====================================================

Obviously, there will be other data in the table (Like title, description, etc.), but I think I probably only need the data above to do comparisons. I will have an array with the id of every classes the user has signed up for.

Sorry if this is a duplicate, I did search and did not see this asked before.

Answer 1

I’m not sure what your primary key is for this table. I’m fairly confident that mine is wrong for you, but it shouldn’t matter for the purpose of your question.

-- student_id and class_id are foreign keys, not shown
create table times (
  student_id integer not null,
  class_id integer not null,
  primary key (student_id, class_id), 

  start_time timestamp not null,
  end_time timestamp not null
);

The first student has an overlap between classes 2 and 3.

insert into times values
(1, 1, '2012-09-01 08:00', '2012-09-01 08:55'),
(1, 2, '2012-09-01 10:00', '2012-09-01 11:55'),
(1, 3, '2012-09-01 11:45', '2012-09-01 12:45');

Second student, no overlaps.

insert into times values
(2, 1, '2012-09-01 08:00', '2012-09-01 08:55'),
(2, 2, '2012-09-01 10:00', '2012-09-01 11:55');

Third student has an overlap between classes 1 and 2.

insert into times values
(3, 1, '2012-09-01 08:00', '2012-09-01 10:00'),
(3, 2, '2012-09-01 09:55', '2012-09-01 11:55'),
(3, 3, '2012-09-01 12:00', '2012-09-01 12:55');

If your platform conformed to SQL-92, you could use the OVERLAPS operator. (If any dbms supported assertions, you could declare the tables in such a way that it wouldn’t be possible to insert overlapping classes.)

select t1.*
from times t1
inner join times t2 on t1.student_id = t2.student_id
       and (t1.class_id <> t2.class_id and t1.start_time <> t2.start_time and t1.end_time <> t2.end_time )
       and (t1.start_time, t1.end_time) overlaps (t2.start_time, t2.end_time);

MySQL doesn’t seem to support that, so you might use the equivalent definition from the SQL standards committee.

select t1.*
from times t1
inner join times t2 on t1.student_id = t2.student_id
       and (t1.class_id <> t2.class_id and t1.start_time <> t2.start_time and t1.end_time <> t2.end_time )
       and (
             (t1.start_time > t2.start_time and not (t1.start_time >= t2.end_time and t1.end_time >= t2.end_time))
             or
             (t2.start_time > t1.start_time and not (t2.start_time >= t1.end_time and t2.end_time >= t1.end_time))
             or
             (t1.start_time = t2.start_time and (t1.end_time <> t2.end_time or t1.end_time = t2.end_time))
       );

I’d expect the join conditions to give you good performance, as long as your tables are indexed carefully. In your case, you’d be selecting rows for a single student, which should be pretty selective. (I didn’t do that in my example code.)