>Hi Hilmar,
>I will start to think about the problem as you suggested. Please continue to work on it. I will keep you posted on my progress. Please keep me posted on your progress.

I can't dedicate much time to this right now, but I am available for any specific questions.

Do you know how program recursive functions? If not, my article in UT Magazine can give you an introduction.

It also occurs to me that, as a shortcut, you keep track of what number combinations are missing. Then, you seek the letter combination that covers the greatest part of the missing combinations. Example (I will use single numbers instead of letter pairs - 'tis the same principle):

You need to cover options:

1 2 3 4 5 6 7 8 9 10

Letter "a" covers options 2 4 6. Let's say your program already decided to use letter "a".

Options still missing: 1 3 5 7 8 9 10

Now, letter b covers options: 1 2 3 (this includes two of the missing options).

And letter c covers options: 5 8 10 (this includes three of the missing options).

You select letter c, and continue the algorithm.

Note: This shortcut should be much faster than the recursion proposed previously - but I believe there is no guarantee that it will always find the most optimal solution. On the other hand, the recursive solution might run for years, or centuries, on large data sets.

Another note: Neither of these algorithms can be solved with simple SQL statements. You will have to program loops.

Hilmar.