Question

I need some help with the following situation:

Very simplified table users_log

+-----------------+-------------------+
| id | account_nr |    email          |
+-----------------+-------------------+
| 1  | 0000001    |  [e-mail]  |
| 2  | 0000001    |  [e-mail]  |
| 3  | 0000002    |  [e-mail]  |
| 4  | 0000003    |  [e-mail]  |
| 5  | 0000002    |  [e-mail]  |
| 6  | 0000001    |[e-mail] |
+-----------------+-------------------+

1. It is allowed for customers to appear multiple times in this table.
2. Customers can change their email addresses over time, old entries are not updated.

As you can see account ‘0000001’ appears 3 times and at some point changed his email address.

I need a query that only returns the results of customers that have never changed their email address. Furthermore I need all single entries, so nothing grouped.

So all entries by customers ‘0000002’ and ‘0000003’ would be the expected result of the query in the example above.

I figured out a way using mutliple loops in a scripting language but wonder if there is a more efficient query I can use hence reducing database load?
This is a very large database and I need the fastest query possible – thank you in advance!

PS: The database structure is this way and I have no way to change anything.

Answer 1

CREATE TABLE users_log (ID INT, account_nr VARCHAR(1000), email VARCHAR(1000))

INSERT INTO users_log VALUES (1,'0000001','[e-mail]');
INSERT INTO users_log VALUES (2,'0000001','[e-mail]');
INSERT INTO users_log VALUES (3,'0000002','[e-mail]');
INSERT INTO users_log VALUES (4,'0000003','[e-mail]');
INSERT INTO users_log VALUES (5,'0000002','[e-mail]');
INSERT INTO users_log VALUES (6,'0000001','[e-mail]');

SELECT account_nr
     , email
  FROM users_log a
 WHERE NOT EXISTS
       (SELECT *
          FROM users_log b
         WHERE a.email      != b.email
           AND a.account_nr = b.account_nr)

Result:

    ACCOUNT_NR  EMAIL
1   0000002     [e-mail]
2   0000002     [e-mail]
3   0000003     [e-mail]