T-SQL: How to Select Values in Value List that are NOT IN the Table?

Question

I have a list of e-mail addresses, some of them are in my table, some of them are not. I want to select all e-mails from that list and whether they are in the table or not.

I can get users whose mail adresses are in the table like this:
SELECT u.* FROM USERS u WHERE u.EMAIL IN ('email1', 'email2', 'email3')

But how can I select values in that list which are not exist in the table?

Moreover, how can I select like this:

E-Mail | Status
email1 | Exist  
email2 | Exist  
email3 | Not Exist  
email4 | Exist  

Thanks in advance.

Solution 1

For SQL Server 2008

SELECT email,
       CASE
         WHEN EXISTS(SELECT *
                     FROM   Users U
                     WHERE  E.email = U.email) THEN 'Exist'
         ELSE 'Not Exist'
       END AS [Status]
FROM   (VALUES('email1'),
              ('email2'),
              ('email3'),
              ('email4')) E(email)  

For previous versions you can do something similar with a derived table UNION ALL-ing the constants.

/*The SELECT list is the same as previously*/
FROM (
SELECT 'email1' UNION ALL
SELECT 'email2' UNION ALL
SELECT 'email3' UNION ALL
SELECT 'email4'
)  E(email)

Or if you want just the non-existing ones (as implied by the title) rather than the exact resultset given in the question, you can simply do this

SELECT email
FROM   (VALUES('email1'),
              ('email2'),
              ('email3'),
              ('email4')) E(email)  
EXCEPT
SELECT email
FROM Users

Solution 2

You need to somehow create a table with these values and then use NOT IN. This can be done with a temporary table, a CTE (Common Table Expression) or a Table Values Constructor (available in SQL-Server 2008):

SELECT email
FROM
    ( VALUES 
        ('email1')
      , ('email2')
      , ('email3')
    ) AS Checking (email)
WHERE email NOT IN 
      ( SELECT email 
        FROM Users
      ) 

The second result can be found with a LEFT JOIN or an EXISTS subquery:

SELECT email
     , CASE WHEN EXISTS ( SELECT * 
                          FROM Users u
                          WHERE u.email = Checking.email
                        ) 
            THEN 'Exists'
            ELSE 'Not exists'
       END AS status 
FROM
    ( VALUES 
        ('email1')
      , ('email2')
      , ('email3')
    ) AS Checking (email)

Solution 3

You should have a table with the list of emails to check. Then do this query:

SELECT E.Email, CASE WHEN U.Email IS NULL THEN 'Not Exists' ELSE 'Exists' END Status
FROM EmailsToCheck E
LEFT JOIN (SELECT DISTINCT Email FROM Users) U
ON E.Email = U.Email

Solution 4

When you do not want to have the emails in the list that are in the database you'll can do the following:

select    u.name
        , u.EMAIL
        , a.emailadres
        , case when a.emailadres is null then 'Not exists'
               else 'Exists'
          end as 'Existence'
from      users u
          left join (          select 'email1' as emailadres
                     union all select 'email2'
                     union all select 'email3') a
            on  a.emailadres = u.EMAIL)

this way you'll get a result like

name | email  | emailadres | existence
-----|--------|------------|----------
NULL | NULL   | [email protected]    | Not exists
Jan  | [email protected] | [email protected]     | Exists

Using the IN or EXISTS operators are more heavy then the left join in this case.

Good luck :)

View more solutions

Comments

[Admin]

That gives the values in the table that are not in the list. The question asks it the other way around: the ones in the list that are not in the table.

[Admin]

With the edit, it still won't be right if Users.email is nullable (and we don't know if it is)

[kubilay]

Thanks, I didn't want to go with long way. Martin's answer did it.

[kubilay]

Thank you, your answer is late but correct, too. I'm glad I didn't have to create a table for values.

[ypercubeᵀᴹ]

@hvd: You are right, we don't know if it is nullable. The query will answer correctly though.

[Martin Smith]

EXISTS will be more efficient than outer join as it can short circuit. Yours can bring back the same address multiple times if there are duplicates in Users. See Subqueries in CASE Expressions for some details as to how EXISTS is handled.

[Admin]

Your second query will, your first won't. WHERE 'a' IN ('b', NULL, 'c') is UNKNOWN rather than FALSE, and NOT UNKNOWN is still UNKNOWN. In other words if any user's email address is NULL, the first query will not return anything.

[ypercubeᵀᴹ]

@hvd: I know but that could be the "correct" behaviour for the OP :) Why else does he have Nulls in the column after all?

[Admin]

Yes, technically, there's a chance that when a user's email address is unknown, that user's happens to be one of the few email addresses in the OP's hardcoded list, and the OP wants to account for the possibility that it is one of those few email addresses. I approximate that chance low enough to discard that possibility.