Structural question (1 Viewer)

[torok]

Hi there,

I'm trying to figure out the best way to do this, and coul duse some input. I currently have a few data columns with primary keys KEYA KEYB KEYC, and a few other data columns with pk KEYA, KEYB, KEYC, KEYD, and a few OTHER columns with pk KEYA, KEYB.

Right now I have three tables (actually, there are even more like this, but I will keep it simple), one table for each set of PKs. This is a complete waste of space, and getting all the KEYA fields to match correctly for joins if I change the name of one of the keys is a pain.

I perform a successive sequence of calculation queries on these tables, creating temp tables along the way, and end up with a final result.

Is there a better way to organize this?

 

[neileg]

I've read this several times and still don't understand what you are doing.

You seem to be suggesting that you have three tables, all of which have a multi field primary key. The fields that you use for these keys have the same names in your tables. OK?

If you have multi field PKs I don't understand how you can match using part of the key, unless you accept that this can be a many to many match. I also don't understand why you need to change the name of your fields.

In fact it's fair to say I don't understand much of this at all!

Can you have another go at explaining. It might help if you set out your table structures and some sample data, together with a better idea of what you want to achieve!

 

[torok]

hmm.... my fault for being cryptic I suppose. You've got it right, except that I never match with part of a key. Just that I have 6 tables, and I do a whole lots of joins.

I think I finally answered my own question, actually. To make this a bit easier there will have to be a lot more joins, is all. Thanks for trying to sort through my bad explanation! Please ignore my question.

I promise if I have any more questions to ask, they will actually make sense.