Search code examples
mysqlsqlcountaggregate-functionsgreatest-n-per-group

Finding the entry with the most occurrences per group


I have the following (simplified) Schema.

CREATE TABLE TEST_Appointment(
    Appointment_id INT AUTO_INCREMENT PRIMARY KEY,
    Property_No INT NOT NULL,
    Property_Type varchar(10) NOT NULL
);

INSERT INTO TEST_Appointment(Property_No, Property_Type) VALUES
    (1, 'House'),
    (1, 'House'),
    (1, 'House'),
    (2, 'Flat'),
    (2, 'Flat'),
    (3, 'Flat'),
    (4, 'House'),
    (5, 'House'),
    (6, 'Studio');

I am trying to write a query to get the properties that have the most appointments in each property type group. An example output would be:

Property_No | Property_Type | Number of Appointments
-----------------------------------------------------
1           | House         | 3
2           | Flat          | 2
6           | Studio        | 1

I have the following query to get the number of appointments per property but I am not sure how to go from there

SELECT Property_No, Property_Type, COUNT(*)
from TEST_Appointment
GROUP BY Property_Type, Property_No;

Solution

  • If you are running MySQL 8.0, you can use aggregation and window functions:

    select *
    from (
        select property_no, property_type, count(*) no_appointments,
            rank() over(partition by property_type order by count(*) desc) rn
        from test_appointment
        group by property_no, property_type
    ) t
    where rn = 1
    

    In earlier versions, one option uses a having clause and a row-limiting correlated subquery:

    select property_no, property_type, count(*) no_appointments
    from test_appointment t
    group by property_no, property_type
    having count(*) = (
        select count(*)
        from test_appointment t1
        where t1.property_type = t.property_type
        group by t1.property_no
        order by count(*) desc
        limit 1
    )
    

    Note that both queries allow ties, if any.