Nando Dessena wrote:

> Dany,
>
> D> IMHO a tool like IBExpert or any DB admin tool should not have the same
> D> goal as a UI client for users. If my application (that my clients use)
> D> finds I missed a PK it will collect the RBD$DB_KEY. If IBExpert did that
> D> too, I would run the risk as a developer to miss these kind of things.
> D> Of course there are ways for a tool to notify, but I prefer my tool
> D> shows more of what is going on under the hood on the server than doing
> D> smart things.
>
> one could say that trying to use all fields to compose the update
> statement (as the reported error message seems to imply) is as
> tentatively smart as trying to use the db_key. Only less effective,
> given the outcome.
>
> There are so many design mistakes you can make designing a database -
> a tool that checks the database structure and hints at possible
> problems would be a good idea. Counting, for the same purpose, on an
> update error while trying to update a row that is identical to another
> one is just silly. Not to mention that, if you issue a delete command
> instead, it'll happily delete the whole bunch. I'm sure that in this
> particular case you'll want your preferred tool to play smarter and
> check the count of affected rows.
>
> Finally, once your preferred tool has notified you that you're fried
> because you forgot to define a PK, how are you going to solve it since
> it doesn't allow you to remove the duplicates easily? I think that
> tools are meant to help me, not just constantly remind me what a fool
> I am. My woman is quite adequate for that. :-)
>

Well... As I said as long as the information given by the tool is
adequate. You are right. But this is very off-topic, no? Sorry about
your woman :)

/D