IBM i SQL to validate Email

So, we have a customer master file that’s been around for years. Users have been keying in customer info and sometimes adding email addresses, sometimes not, sometimes entering nonsense, or even just text that says things like “this chap has no email so don’t email him”.

Yes, it’s a filthy dirty file filled with grubby email addresses. #grubby

It’s time to get cleaning….

Cleanup the email addresses in any database

Look for any email addresses that are NOT in the format somebody@something.extension

Luckily it is super easy to do a quick and basic email validity check using SQL:

SELECT * FROM 'file-name' WHERE 'email-field' NOT LIKE '%@%.%'

This is a basic selection that will check for really obviously malformed email address. It does not check if an email address is actually *real* or not.

We can add some more checks to this without getting into regular expression madness:

select * from 'file-name'
 where 'email-field' not like '%@%.%' -- Must be someone@something.somewhere
or email not like '%@%.%' -- Must contain at least one @ and one
or email like '%..%' -- Cannot have two periods in a row
or email like '%@%@%' -- Cannot have two @ anywhere
or email like '%.@%' or email like '%@.%' -- Cannot have @ and . next to each other
or email like '%.cm' or email like '%.co' -- Unlikely. Probably typos 
or email like '%.or' or email like '%.ne' -- Missing last letter

Tweak it to your heart’s delight but it’s a simple and easy way to validate all email addresses within any IBM i table.

I’ve used this technique dozens of times and it’s very easy to visually pull a list of entries that may be malformed. However, if you want to programmatically validate an email address then we can look at use an SQL REGEX to do that validation or write a high-level script/program to do it.

Since I don’t have a good RPG ILE open-source program to do this – let’s look at a REGEX example: