Question

In Microsoft SQL Server 2005 or above, I would like to get the first row, and if there is no matching row, then return a row with default values.

SELECT TOP 1 ID,Name
FROM TableName
UNION ALL 
SELECT 0,''
ORDER BY ID DESC

This works, except that it returns two rows if there is data in the table, and 1 row if not.
I’d like it to always return 1 row.
I think it has something to do with EXISTS, but I’m not sure.
It would be something like:

SELECT TOP 1 * FROM Contact 
WHERE EXISTS(select * from contact)

But if not EXISTS, then SELECT 0,”

Answer 1

What happens when the table is very full and you might want to specify which row of your top 1 to get, such as the first name? OMG Ponies’ query will return the wrong answer in that case if you just change the ORDER BY clause. His query also costs about 8% more CPU than this modification (though it has equal reads)

SELECT TOP 1 *
FROM (
   SELECT TOP 1 ID,Name
   FROM TableName
   ORDER BY Name
   UNION ALL
   SELECT 0,''
) X
ORDER BY ID DESC

The difference is that the inner query has a TOP 1 also, and which TOP 1 can be specified there (as shown).

Just for fun, this is another way to do it which performs very closely to the above query (-15ms to +30ms). While it’s more complicated than necessary for such a simple query, it demonstrates a technique that I don’t see other SQL folks using very often.

SELECT
   ID = Coalesce(T.ID, 0),
   Name = Coalesce(T.Name, '')
FROM
   (SELECT 1) X (Num)
   LEFT JOIN (
      SELECT TOP 1 ID, Name
      FROM TableName
      ORDER BY ID DESC
   ) T ON 1 = 1 -- effective cross join but does not limit rows in the first table