Re: Enforcing serial uniqueness? - Mailing list pgsql-general

From Csaba Nagy
Subject Re: Enforcing serial uniqueness?
Date
Msg-id 1143035283.23597.37.camel@coppola.muc.ecircle.de
Whole thread Raw
In response to Re: Enforcing serial uniqueness?  (Steven Brown <swbrown@ucsd.edu>)
List pgsql-general
> Gaps are fine.  All I want is safe uniqueness.  What is an issue for me
> is a user having INSERT permission being able to shut down all INSERTs
> from everyone else until someone manually figures out what happened and
> fixes it, ditto for UPDATE permission on a sequence (which they need in
> order to use nextval so they know what id the row they inserted will
> have, right?), which seems extremely dangerous to me.

No doubt that you will be able to solve your problem by the permission
system somehow, but I wonder how stable a system can be if you allow
users to insert directly to the data base ?

If you would have a set of fixed/approved SQLs which are controlled by
an application (and the application in turn controlled by responsible
developers you trust), and users can only access data through that
application, these kind of problems won't occur at all...

Cheers,
Csaba.



pgsql-general by date:

Previous
From: Steven Brown
Date:
Subject: Re: Enforcing serial uniqueness?
Next
From: "Jim Buttafuoco"
Date:
Subject: Re: Enforcing serial uniqueness?