The only times 1 to 1 relations are really usefull are if you store BLOBS
(OLE/memo things) and don't want to retrive them all the time (and you're to
*lazy* to use select a,b,c when you don't need that field) and/or the rare
occassions where you need more than 127 (255) fields in a table
My guess is that you've set up something like
Table LOANS
LOANID (Autonumber)
LOANNUMBER (Text/Long Int)
Table LOANDETAILS
LOANID (long int - fk)
.....
and thus (since you've learned that "dumb" keys is *the* way to make
relational db's - and you didn't need a relation (?!) - you went by the
book)
you ought to merge you tables into one....
it would however make sense to use a "dumb" key to link to the customer
table (1 -> many) as a person might have several loans ....
Pieter
SCHNYDES said:
Actually having the one to one has been useful, in helping us keep
duplicate values from the tables. I work for a mortgage company, and the
loans we track we track by loan number. One Loan(unique) and it's values.
Any suggestions on cleaning up my table structure? I refuse to use an
autonumber, it must go by Loan number to prove accuracy.forcing referential integrity. Often, the relationships disappear, and
sometimes cause my tables to be non-uniform. Why is this and how can I
prevent it?