Count of Items in one row but not in another

Question

I have an access table that has in one column a list of groups and then another column that has a list of products. I would like to run a report to show how many groups are in all products, how many groups are at least 1 of the products. I have as many as 6 products in a given table and over 5,000 unique groups.
Below is an example of what I am looking for:

Table:

Group | Product  
AAAA  | 123456  
AAAA  | 234578  
AAAA  | 456789  
AAAA  | 789012  
BBBB  | 123456  
BBBB  | 234578  
BBBB  | 456789  
CCCC  | 123456  
CCCC  | 234578  
CCCC  | 456789  
CCCC  | 789012  

Expected Results:

Product 123456 and 234578 and 456789 and 789012 has 2 groups  
Product 123456 or 234578 or 456789 or 789012 or has 3 groups  

I tried the below query but it is only for when it is a product or another product not when in both. I am looking for when a group is in product x and product y and product z but not in product a.

  SELECT group  
  FROM GroupProducts  
  WHERE   
         product in (’123456’,’2345678’,’456789’)  
         and product not in (’789012’)

Answer 1

Try to use Count and group by Product :

SELECT Count(group), Product
FROM GroupProducts
WHERE
product in ('123456','2345678','456789')
and product not in ('789012')
GROUP BY product

This assumes your table doesn't have duplicates, else you'll have to filter them out first

Answer 2

I would like to run a report to show how many groups are in all products

Assuming no duplicate records, the following query should return the set of all groups which are associated with all products in the dataset:

select t.group from YourTable t group by t.group
having count(t.group) = 
(
    select count(*) from (select distinct u.product from YourTable u)
)

You can count these groups by simply enclosing the above with a select count(*) query:

select count(*) from
(
    select t.group from YourTable t group by t.group
    having count(t.group) = 
    (
        select count(*) from (select distinct u.product from YourTable u)
    )
)

how many groups are at least 1 of the products

This one is significantly easier, since you can simply select the number of distinct groups in the dataset:

select distinct t.group from YourTable t

Which can then be counted in the same manner as previously described:

select count(*) from 
(
    select distinct t.group from YourTable t
)

---

In all of the above examples, replace all instances of YourTable with the name of your table.