Re: NLS character settings

> Hi Robert.
>
> There are two points to note here:
> 1. Character set conversions do not guarantee a round trip.
> 2. You are dealing with characters not code point values.
>
> I do not know what character 0x9E represents in either character set but
> presumable [based on your results] it represents a different character in
> each of the chosen character sets. On the client side you have chosen
> EE8MSWIN1250. When the character is passed to the server, the code-point
> value 0x9E is not simply passed to the server. Because they are different
> character sets, Oracle will first convert the character to a Unicode
> code-point [so the doco says] and then attempt to convert this Unicode
> character to the target character set [in your example WE8DEC]. In your
> example, the Unicode character is not able to be represented in WE8DEC so
> the replacement character [i.e. a '?'] is used.
>
> When you use the same characterset on the client as on the server, Oracle
> does not perform any conversion [or validation for that matter].
>
> Hope this helps.
>
> Cheers Chris

Hello Chris,
character 0x9E is 'ž' in slovak character set. it is 'z' with a little sign above.

1. I did not know it.
2. that`s what I thought. As I understood, the conversion guarantees that oracle gives the same graphical representation of character if the same client-server combination of NLS settings are used. OK seems it is not always true.

Just one more qestion.
Is there any relationship between OS(windows) national settings, and oracle client NLS settings. For example if I use english windows regional settings and Oracle NLS settings are set to Slovak (EE8WIN1250), or any different 8bit Oracle NLS-OS national combination. Can you explain me please what kind of conversions are going on background.

Regards
Robert Received on Thu Mar 04 2004 - 01:46:10 CST