Question

I’m racking my brain trying to work out what group by and/or distinct and count() formation help with this but have currently drawn a blank.

Consider the table:

PersonId | PlaceName
---------+----------
    1    |   NULL
    1    |   NULL
    2    |   NULL
    2    | Las Vegas
    3    | London
    4    |   NULL
    4    |   NULL
    4    |   NULL

I’m looking for people who haven’t bothered to fill in “place names”, so I want my output to look like this:

PersonId
--------
    1
    4

I’ll actually be joining some other tables in to pull out the info on each “miscreant” person, but the crux of my problem is the one above.

Answer 1

Use the following query:

SELECT PersonId
  FROM TheTable
 GROUP BY PersonId
 HAVING COUNT(PlaceName) = 0

The COUNT() aggregate function ignores NULLs, so the right result should be returned.