I have almost no programming knowledge, so I'm a little (probably a lot) lost on the basics of joins. I have 7 tables which I joined at the beginning of my program.

The 1st table is joined to 3 tables (2 use the primary key, 1 does not). Then the other 3 tables are joined to each other by whatever field the have in common (sometimes primary key, sometimes not).

1. Should I be putting all of the joins at the beginning?

2. Some of the tables are joined by the primary key; some are not because they have a different field in common. Is that important?

3. Can I join A to B, A to C, A to D. Then join B to E, B to F, F to G?

3. This program works sometimes (very slow) & other times it just runs & runs until I kill it. I'm assuming that's due to the way I wrote it.

4. Where can I learn this stuff? I looked up all of the join documentation I could find, but it didn't really tell me if I could put them all at the beginning; if primary keys mattered; if I needed to do some kind of indexing; if I could join A to B, and then B to C, and then D to B, etc, etc, etc.

Thanks.

Corky
So many questions, and on such a complex issue. It is useful for us to know a lot more about your system, type of DB etc. Always helps.

But in general:
1. It is normal to create a JOIN structure at the beginning of the program. Only Tables with referenced fields in the report will be accessed. However, it is worth looking at a cluster join, which is an mfd, rather than issuing JOIN syntax everywhere.
2. This is down to your DBA. It is not the Primary Key that is important but indexing on the required fields that you are joining to. Without this performance can be wickedly slow.
3. When you join A to B, B becomes part of A, so a subsequent JOIN to B references file A, as B is now part of the structure A. Same applies to F.
4. See 2.
5. The documentation does assume a certain level of knowledge. Talk to your local IB rep to get some training. This can be hugely beneficial.