Graphic problem!

[Biel]

Hello!


I have created a table in which I have a "multiple value field".
That means that a client may have "A" or "B" or "C", as well he may have "A and C", "A and B", "A, B and C", etc.
When I make a chart from that, to know how many of them have A, B and C, the ones that have more than one are separated. The ones that have "A and B" for an example, have a slice just for themselves, instead of being included in "A" and in "B".
How do I solve that?
Thank you.

[orange]

Most will recommend you do NOT use multivalued fields.

Normalize your tables
For free normalization video tutorial see
http://www.lynda.com/Access-2003-tut...n/13150-4.html

[boblarson]

I agree with orange. I would say that multi-valued fields can be a real pain to actually use in practice because of several limitations such as this. The benefit of a pretty user interface is quickly overshadowed by real problems in pulling the data back out. I would highly suggest going to using a separate table for the selected items and either use a subform or a listbox with code to have multi-select actually be available for entry, but it will simplify things for when you need to report on the data.

[hertfordkc]

Hello!
I have created a table in which I have a "multiple value field".
That means that a client may have "A" or "B" or "C", as well he may have "A and C", "A and B", "A, B and C", etc.
When I make a chart from that, to know how many of them have A, B and C, the ones that have more than one are separated. The ones that have "A and B" for an example, have a slice just for themselves, instead of being included in "A" and in "B".
How do I solve that?
Thank you.

I agree with Orange and BobLarson's comments. However, I understand that your question has to do with how to assemble the data to create a report. To that aspect, do you agree that entries with a single response and entries with all responses are not a problem? Your concern is how to count responses of more than one value but less than all? If this is the case, I don't have an immediate answer concerning how to do this with a query or queries. However, with the working copy of the dataset, one could loop through all of the data, counting the combinations as they were removed from the dataset. I expect that this can be accomplished with SQL, but I can't help you there.