How can I select all the rows which do not share a column value with another row which is null?

Question

I know that title is word vomit, but I wasn't sure how to say this more concisely.

I have a very large table with columns A and B. This table is supposed to have an unusual property: for every value of column A, there should be at least one row where the value of B is null. However, this table has some anomalous rows which do NOT abide by this property and I would like to identify them so that I can clean up the table.

id  A   B
----------
1   c   e
2   c   f
3   c   NULL
4   d   e
5   d   f
6   d   g

In this example, the rows where column A is c satisfy the criteria because row 3's B value is NULL. However, the rows where column A is d do not satisfy the criteria because there is no row where column A is d and column B is null. I would like to run a query which returns d.

Answer 1

you can try this

SELECT DISTINCT A from Test where A not in (
select A from Test where B is null );

Answer 2

You could also use aggregation:

select d
from t
group by d
having count(*) = count(b);