Unless I am missing something, <unique> only makes the individual field unique. What I need is the combination to be unique. In other words...
If SESSION_ID = 10 SECTION_ID = 15 ACTIVITY_ID = 25 then 10, 15, and 25 can't be again used together.. but SESSION_ID = 10 SECTION_ID = 15 ACTIVITY_ID = 30 Would work because even though SESSION_ID and SECTION_ID have been 10 and 15 before, they have not been in combination with 30. I could just make ID the key but then the following would be allowed ID = 1 SESSION_ID = 10 SECTION_ID = 15 ACTIVITY_ID = 25 ID = 2 SESSION_ID = 10 SECTION_ID = 15 ACTIVITY_ID = 25 Etc. etc.... Jeff -----Original Message----- From: Dave Newton [mailto:[EMAIL PROTECTED] Sent: Wednesday, March 03, 2004 9:10 PM To: Apache Torque Users List Subject: RE: autoincrement of non primary key On Wed, 2004-03-03 at 20:52, Jeff Cox wrote: > Dale, > First let me say I really appreciate you responding to my question. > Secondly, I would agree with you and in fact considered that possibility > early on in the design, but the truth is I need any given combination of > (SESSION_ID, SECTION_ID, ACTIVITY_ID) to be unique. I don't know of > another way to enforce this in the schema, except by making them the > primary key. If there is another way to do this my problem would > certainly be solved... The <unique> stuff doesn't work? Wait, you need _any_combination_ of those three values to be unique? I don't get it. If any combination of those needs to be unique then you only need one number. Dave --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
smime.p7s
Description: S/MIME cryptographic signature
