|
Posted by Tony Rogerson on 09/30/59 11:47
>I think Oracle got it right. It should not matter if you use MYTABLE,
> MyTable, or mytable in your SQL statement FROM clause. The database
> should recognize the table name. But data should be stored exactly as
> entered so that mike, Mike, and MIKE should be different. There are
As for data Oracle has it completely wrong, collation should reflect the
real world if a salesman shouts out AKA do they mean aka, Aka, AKa, AKA or
what???? You should not have to rely on doing this data verification
yourself, adding constraints to check case etc... is just making up for a
poor initial decision to go with case sensitivity.
This is the biggest pain in the arse in the business intelligence and
reporting area.
Case Insensitive is how the real world works and that should be the default
for any database 'period'.
As for Object names it shouldn't matter but case insenitive promotes poor
programming practice.
Consider portability to other products and systems.....
IMHO....
--
Tony Rogerson
SQL Server MVP
http://sqlserverfaq.com - free video tutorials
"Mark D Powell" <Mark.Powell@eds.com> wrote in message
news:1147449278.247001.136270@i39g2000cwa.googlegroups.com...
>I think Oracle got it right. It should not matter if you use MYTABLE,
> MyTable, or mytable in your SQL statement FROM clause. The database
> should recognize the table name. But data should be stored exactly as
> entered so that mike, Mike, and MIKE should be different. There are
> easy ways to handle forcing the data into a standard storage format via
> column level constraints, table triggers, and single row functions like
> UPPER, LOWER, and INITCAP. And there are ways to perform case
> insensative compares where needed or desired.
>
> Standarding data entry so that you do not end up with what amounts to
> duplicates due to variation of entry has been an application issue
> since computers were invented. Example: Johnson Inc. vs Johnson
> Incorporated. Allowing the CAPLOCK key to add additional variables to
> the mix is just asking for data integrity problems. Ever column of
> every table in the application should have data entry rules. Then you
> do not end up having to know or figure out if you are looking for MIKE,
> mike, or Mike.
>
> IMHO -- Mark D Powell --
>
Navigation:
[Reply to this message]
|