Oracle FAQ | Your Portal to the Oracle Knowledge Grid |
Home -> Community -> Usenet -> c.d.o.misc -> Re: Primary Key of type string ?
Guys,
I'm an MS SQL guy and wanted to verify something before I said
anything (at work).
If the only business requirement in an Oracle 10 db is that a BO (biz
object) has a unique identifier, and this identifier is not created
using data from said BO, would it ever make sense to use char data for
this id field?
The procs used to create this BO generate this UI from a numeric type
value randomly created (no chars or padding is ever used, so it stays
numeric) then it converts it to char type data before writing it to
the db. This db is riddled with this type of identifier in various
tables which house various parts of this BO, all having there own
unique id'ers created this same way, none of which use BO data in this
unique column. I do not foresee a performance hit to speak of upon
initial creation of these BO's, as there may only be 1000 or so
created every hour, but the application will reference this data
heavily across joined tables and canned reports, as well as adhoc
queries. IYO, would I be making a mountain out of a molehill by
mentioning this to the developers as a potential performance lag? I
would never do this in MS SQL unless, as mentioned above, rules
required it (even then I would be testing performance vs. a
workaround).
Please advise and thank you in advance.
Oriane wrote:
> Hello,
>
> after reading the thread, I must admit there is no standard choice !!!!!
>
> Regards
Received on Fri Jun 08 2007 - 09:03:44 CDT