I want to thank all of you for your advices. Before I could read your
last post I have worked out a solution. As you

previously said, table 2 (CHOICE), is ill-prepared for what I want to
do, so I had to create a union query to solve this.

This query would be:

SELECT Product_Name as PName, 1 as Order
FROM CHOICE, PRODUCTS
WHERE FIELD_1 = PRODUCTS.ID

UNION ALL

SELECT Product_Name as PName, 2 as Order
FROM CHOICE, PRODUCTS
WHERE FIELD_2 = PRODUCTS.ID

UNION ALL

SELECT Product_Name as PName, 3 as Order
FROM CHOICE, PRODUCTS
WHERE FIELD_3 = PRODUCTS.ID;

This query (lets name it ARRANGED_DATA) contains now the names of the
brands along with the order of appearance. (It's very

important the ALL keyword, which indicates that the query must include
duplicated values.)

Having done this, is easy to set up another query to obtain the final
results:

SELECT PName  AS Products,
-sum(Order=1) AS Place_1,
-sum(Order=2) AS Place_2,
-sum(Order=3) AS Place_3
FROM ARRANGED_DATA
GROUP BY PName;

And that's all folks! Yes, I know it's very nasty the
-sum(BOOLEAN_EXPRESSION), but right now I can't figure out something

neater as a replacement for it (i.e. to count the YES occurrences in
the boolean expression). If you know something better,

I'd like to know about it.

I guess the solution you propose is better and I'll use it instead of
this one I listed here.

Thanks again for your time and concern.

Sergio

Duane Hookom ha escrito: