From: Laurenz Albe on
Ben <balvey(a)comcast.net> wrote:
> I'm re-reading the Globalization Guide again, it's still confusing.
>
> After reading this it sounds like Oracle is saying that conversion is
> a bad thing and that the database character set should be set the same
> as the client. This is the reason why there is so much confusion, at
> least on my part, on character sets.

It is a good thing that you are spending so much thought on the matter
before you create your database - once you have created your database,
it is hard to correct any design flaws you have made.

I agree with you that reading what you quoted from Oracle documentation
sounds like they are all FOR having database character set and client
character set identical. And true, other parts read like they are dead
against it.

The truth of the matter is that there is no 100% correct solution without
drawbacks. You'll have to weigh the pros and cons.

What speaks for having client and server use the same encoding,
e.g. WE8MSWIN1252:

- No conversion overhead.

What speaks against it:

- No check if valid bytes are entered (can lead to garbage in database).
- You can only store Windows-1252 characters in the database.

I don't know, but I guess that conversion on characters in messages between
client and server will not be the performance bottleneck of your database;
usually it is disk I/O.

Ask yourself questions like:

- Is there a chance that, at some point, outlandish characters should be
stored in a text field that was not originally designed for it?

- Can I be absolutely sure that no client will attempt to store characters
in a wrong encoding (e.g., you have to import table data from a
different company)?

Yours,
Laurenz Albe
From: sybrandb on
On Tue, 28 Aug 2007 10:54:57 -0700, Ben <balvey(a)comcast.net> wrote:

>After reading this it sounds like Oracle is saying that conversion is
>a bad thing and that the database character set should be set the same
>as the client.

That is EXACTLY what Oracle is NOT (I REPEAT: NOT) saying.

Oracle SAYS:
THE DATABASE CHARACTERSET SHOULD BE THE SAME AS THE CHARACTERSET OF
THE O/S.

WHAT IS SO DIFFICULT TO UNDERSTAND ABOUT THAT?

--
Sybrand Bakker
Senior Oracle DBA
From: sybrandb on
On 29 Aug 2007 08:03:45 GMT, Laurenz Albe <invite(a)spam.to.invalid>
wrote:

>
>I agree with you that reading what you quoted from Oracle documentation
>sounds like they are all FOR having database character set and client
>character set identical. And true, other parts read like they are dead
>against it.

NONSENSE, UTTER NONSENSE. ORACLE STATES DATABASE CHARACTERSET SHOULD
MATCH O/S CHARACTERSET
>
>The truth of the matter is that there is no 100% correct solution without
>drawbacks. You'll have to weigh the pros and cons.
>
>What speaks for having client and server use the same encoding,
>e.g. WE8MSWIN1252:
>
>- No conversion overhead.
>
>What speaks against it:
>
>- No check if valid bytes are entered (can lead to garbage in database).
AGAIN NONSENSE. THERE WON'T BE ANY 'GARBAGE' IN THE DATABASE!!!
>- You can only store Windows-1252 characters in the database.
>

>I don't know, but I guess that conversion on characters in messages between
>client and server will not be the performance bottleneck of your database;
>usually it is disk I/O.
>
>Ask yourself questions like:
>
>- Is there a chance that, at some point, outlandish characters should be
> stored in a text field that was not originally designed for it?
>
>- Can I be absolutely sure that no client will attempt to store characters
> in a wrong encoding (e.g., you have to import table data from a
> different company)?
PROVIDED THE CLIENT CHARACTERSET DURING IMPORT WAS SET CORRECTLY, THAT
DATA IS GOING TO BE C O N V E R T E D AUTOMATICALLY!!!



--
Sybrand Bakker
Senior Oracle DBA
From: Laurenz Albe on
sybrandb(a)hccnet.nl wrote:
> NONSENSE, UTTER NONSENSE. ORACLE STATES DATABASE CHARACTERSET SHOULD
> MATCH O/S CHARACTERSET

You are not only rude, but also clueless.

Read, for example,
http://www.oracle.com/technology/pub/columns/trute_unicode.html
or Metalink Note 333489.1

"Oracle recommends using Unicode for all new system deployment."

Laurenz Albe
From: sybrandb on
On 29 Aug 2007 09:26:09 GMT, Laurenz Albe <invite(a)spam.to.invalid>
wrote:

>You are not only rude, but also clueless.

Sucker.

How often do you post anything useful?

--
Sybrand Bakker
Senior Oracle DBA