I want to build an oracle function, in a 10g Unicode database, that
accepts a varchar, of any length, and returns a translation. It will
translate all found DOS NL to UNIX NL, or UNIX NL to DOS NL. It has
to be bullet proof. I'm off to give this a try with the regular
express package, but I'd like to get any input from the Oracle Jones.
Thank you.

(David Fitzjarrell??? where are you? Punish me please! :-)

On Thu, 13 Jan 2011 12:52:30 -0800, oldyork90 wrote:

Let me try and take a shot: you would need to replace \n\r (DOS standard)
with \n (Unix standard). You mentioned Oracle 10g, which supports regular
expressions. That shouldn't be very hard to do.



--
http://mgogala.byethost5.com

On Jan 15, 1:51*pm, Mladen Gogala <gogala.mla... (AT) gmail (DOT) com> wrote:

If he were a programmer type, he could also find the C sources for
DOS2UNIX.c and UNIX2DOS.c and create an external procedure that does
the work for him without trying to re-invent the wheel...

Just a thought...

Mladen:

#t Let me try and take a shot: you would need to replace \n\r (DOS
standard) with \n (Unix standard). You mentioned Oracle 10g, which
supports regular expressions. That shouldn't be very hard to do.

Why do it inside a database call what the operating systems provide
already?

On Sat, 15 Jan 2011 17:45:53 -0800, John Hurley wrote:

Because that's what he asked about. I agree with you, I normally don't do
it within the database.



--
http://mgogala.byethost5.com

On Jan 16, 12:23*pm, Mladen Gogala <gogala.mla... (AT) gmail (DOT) com> wrote:
Exactly. I also agree with you Mladen. What he should really be
doing is to trim this sort of thing before ever storing it in the
database. However, for every "should be" there are many other "but in
THIS case..."

Sometimes ya just can't win.