Here's a quote from the Oracle reference guide, page 27:

"There are ways to find information from the database regardless of
whether the data is in uppercase or lowercase, but these methods impose
an unnecessary burden. With few exceptions, such as legal text or
form-letter paragraphs, it is much easier to store data in the database
in uppercase. It makes queries easier and provides a more consistent
appearance on reports. When and if some of this data needs to be put
into lowercase, or mixed uppercase and lowercase (such as the name and
address on a letter), then the Oracle functions that perform the
conversion can be invoked. It will be less trouble overall, and less
confusing, to store and report data in uppercase."

Is it just me, or is this advice profoundly retarded? I'd love to see
what those Oracle functions would do with a name like MacDonald, or a
company name like eBay. Especially in the age of StudlyCaps, I don't
think the correct capitalization of proper names is something Oracle
can just automagically figure out.

Laszlo

chaoslight (AT) gmail (DOT) com wrote:

Hmmm let me see:
- No version
- When I take the Oracle Reference Manual (from
http://tahiti.oracle.com) version 9.2 and go to page 27, I'm still in
the index.

So, provided function based indexes do exist, and are now included in
Oracle Standard:

WTF, why do you post this utter drivel?
Don't you have more important issues to attend to?


--
Sybrand Bakker
Senior Oracle DBA

My $0.02 (USD)...

If the format of the raw data is important (like a customer name),
store it that way. Don't 'cleanse' it, don't convert it. If I want to
be referred to as $*#dwf*(@#Q$* (or the DBA formerly known as Prince),
that is how you should communicate with me. As a customer, I don't care
about Codd, 3NF, data cleansing, etc. I went to great pains to change
my name, I demand that you respect it.

If you need to query on it, you have several options. You can store it
in another column in a cleansed/converted format. You can create
function-based indexes, views with functions, etc.

This is an issue for the business and users to decide. Before you apply
a theory or standard to a business situation, always ask if it makes
sense.

Regards,
Daniel Fink

Daniel Fink wrote:

ROFL!

Comments embedded.
chaoslight (AT) gmail (DOT) com wrote:
Really ... which guide would that be? I find no such 'quote' on any
page 27 of any Oracle documentation at tahiti.oracle.com.

Obviously this is a fabrication intended to do nothing more than
provoke an argument, as it does not resemble, even remotely, the
writing style found in the Oracle documentation. Why would any
reliable database vendor make such ludicrous statements in their
reference manuals? Never have I seen any such 'recommendation' from
Oracle; possibly you were perusing the Mysql documentation or your
class text and thus attributed this incorrectly.

Possibly it's the original poster who's 'retarded' by posting such
nonsense in hopes of a flame war....

The same thing such functions would do in SQL Server, Sybase, Informix,
Mysql or DB2:

SQL> select upper('MacDonald') from dual;

UPPER('MA
---------
MACDONALD

SQL> select lower('MacDonald') from dual;

LOWER('MA
---------
macdonald

SQL> select initcap('MacDonald') from dual;

INITCAP('
---------
Macdonald

SQL> select upper('eBay') from dual;

UPPE
----
EBAY

SQL> select lower('eBay') from dual;

LOWE
----
ebay

SQL> select initcap('eBay') from dual;

INIT
----
Ebay


The operative phrase in that last statement is 'I don't think', which,
apparently, you don't do very well. Note the last output from each
example above, using the initcap() function; Oracle has no problem
setting such case nor does it have any problem recognising it in the
database.


David Fitzjarrell

fitzjarrell (AT) cox (DOT) net írta:
Oracle9i: The Complete Reference, by "Kevin Loney, George Koch, And the
Experts at TUSC".

..... heh.

Are you always this shrill?

You bore me now.

Laszlo

chaosli... (AT) gmail (DOT) com wrote:
Oh, so that should have been here is a quote from a book written about
Oracle.

Obviously Oracle is the only product ever to have had a book written
about it that is not completely accurate and contains questionable
advice that cannot be followed without questioning or thinking for
yourself.

Perhaps you would be better off emailing the publishers or the authors
with your questions and concerns.

Still, it comes from a fairly respected publication. Perhaps the
"advice" given within that book could be debated. That's certainly not
beyond the remit of this newsgroup, to my knowledge.

Why? Maybe the OP has already done so yet is also interested in the
opinion of some of the self-confessed "experts" around here.

mjbox01 (AT) gmail (DOT) com írta:
I am not an Oracle expert. I am merely competent. I felt that the
advice written in that book was stupid, but I thought that it was quite
possible that I had overlooked something. This seemed like the best
place to ask about it.

Perhaps you are right. But it seems like the publishers/authors of a
book would not be the best people to ask for an objective opinion of
their work.

Laszlo

In article <1137596468.281193.53290 (AT) o13g2000cwo (DOT) googlegroups.com>,
says...
I think it was a perfectly reasonable place to bring up the question.
However, you should have taken the time to identify the name of the
aource you were quoting in the first place - phrasing it as "Here's a
quote from the Oracle reference guide, page 27:" does imply to the group
that it is material from Oracle.

--

jeremy