Validate email addresses in Mysql

Question

This query creates a mysql view that captures bad email address formats in one table. So if a row is inserted in that has rtrrg.com as a email it will be recorded in the view. My question is, how do I make the view track more than one table. A second table.

The SQL

CREATE VIEW `invalid_emails` AS 
  select `table_with_email_column`.`email` AS `invalidemail` 
    from `table_with_email_column` 
   where ((locate(_latin1'', ltrim(rtrim(`table_with_email_column`.`email`))) <> 0) 
      or (left(ltrim(`table_with_email_column`.`email`), 1) = _latin1'@') 
      or (right(rtrim(`table_with_email_column`.`email`), 1) = _latin1'.') 
      or ((locate(_latin1'.', `table_with_email_column`.`email`,locate(_latin1'@', `table_with_email_column`.`email`)) -  locate(_latin1'@', `table_with_email_column`.`email`)) <= 1) 
      or ((length(ltrim(rtrim(`table_with_email_column`.`email`))) -  length(replace(ltrim(rtrim(`table_with_email_column`.`email`)), _latin1'@', _latin1''))) <> 1) 
      or (locate(_latin1'.', reverse(ltrim(rtrim(`table_with_email_column`.`email`)))) < 3) 
      or (locate(_latin1'.@', `table_with_email_column`.`email`) <> 0) 
      or (locate(_latin1'..', `table_with_email_column`.`email`) <> 0));

Solution 1

You can use a pure SELECT to validate Email Addresses:

SELECT * FROM `users` WHERE `email` NOT REGEXP '^[^@]+@[^@]+\.[^@]{2,}$';

And now for your question of tracking multiple tables, you can use comma seperated table names right?

SELECT * FROM `users`, `customers`, `clients`
         WHERE `email` NOT REGEXP "^[a-zA-Z0-9][a-zA-Z0-9.!#$%&'*+-/=?^_`{|}~]*?[a-zA-Z0-9._-]?@[a-zA-Z0-9][a-zA-Z0-9._-]*?[a-zA-Z0-9]?\\.[a-zA-Z]{2,63}$";

Solution 2

For the proper email validation, you can use this regex as bellow:

SELECT
    *
FROM
    `school`
WHERE
    `email` NOT REGEXP '^[a-zA-Z0-9][a-zA-Z0-9._-]*[a-zA-Z0-9._-]@[a-zA-Z0-9][a-zA-Z0-9._-]*[a-zA-Z0-9]\\.[a-zA-Z]{2,63}$';

Solution 3

Simple SELECT statement is sufficient, for example:

 SELECT * FROM user WHERE email NOT 
 REGEXP '^[a-zA-Z0-9][+a-zA-Z0-9._-]*@[a-zA-Z0-9][a-zA-Z0-9._-]*[a-zA-Z0-9]*\\.[a-zA-Z]{2,4}$'

This query handles the Gmail addresses with + sign and addresses where the host is a single letter.

Solution 4

You can use a UNION in the VIEW but then you have to repeat all the WHERE statement which gives you redundant code. So you would make a helper VIEW that makes you a UNION and then apply the WHERE clause.

Demo here: SQL Fiddle Demo.

That would apply to your SQL somehow like this (untested);

CREATE VIEW `invalid_emails_helper` AS
  select `table_with_email_column`.`email` AS `invalidemail` 
    from `table_with_email_column` 
union
  select `table_with_email_column`.`email` 
    from `second_table_with_email_column` 

CREATE VIEW `invalid_emails` AS 
  select `invalidemail` as `email`
    from `invalid_emails_helper` as `table_with_email_column` 
   where ((locate(_latin1'', ltrim(rtrim(`table_with_email_column`.`email`))) <> 0) 
      or (left(ltrim(`table_with_email_column`.`email`), 1) = _latin1'@') 
      or (right(rtrim(`table_with_email_column`.`email`), 1) = _latin1'.') 
      or ((locate(_latin1'.', `table_with_email_column`.`email`,locate(_latin1'@', `table_with_email_column`.`email`)) -  locate(_latin1'@', `table_with_email_column`.`email`)) <= 1) 
      or ((length(ltrim(rtrim(`table_with_email_column`.`email`))) -  length(replace(ltrim(rtrim(`table_with_email_column`.`email`)), _latin1'@', _latin1''))) <> 1) 
      or (locate(_latin1'.', reverse(ltrim(rtrim(`table_with_email_column`.`email`)))) < 3) 
      or (locate(_latin1'.@', `table_with_email_column`.`email`) <> 0) 
      or (locate(_latin1'..', `table_with_email_column`.`email`) <> 0));

And yes, the query to check the e-mail address using a regex as can easily found everywhere in the internet simplifies it further.

View more solutions