How to create unbuffered file descriptor in ksh.

[Richard Ray]

How do I create a new file descriptor in ksh and make it unbuffered?

[harry d brown jr]

Richard, ksh allows you to create files, but you can't control wether it's buffered or not. Can you describe a little more as to what you are trying to do?

live free or die
harry

[Santosh Nair_1]

I believe Harry is correct, ksh does not give you this ability. Perl however does...any possibility of using Perl instead?

-Santosh

[Richard Ray]

Thanks guys. That's what I thought. My script uses isql to do a sql query in Sybase. There is another process waiting for the results. The query takes a couple of hours so I would like the search results returned as it's found not buffered and delivered in chunks. I guess Perl it is.

[Santosh Nair_1]

If you're going to use Perl with sybase, you might also want to look into SybPerl, which is perl with sybase extensions. Alternatively you could look at DBI/DBD which is a more abstract way of accessing Sybase within Perl.

Hope this helps.

-Santosh

[harry d brown jr]

Santosh is correct, that the perl DB extensions are great. It makes "scripting" a lot easier. There are at least "extensions" for Informix, Oracle, and I believe mysql.

live free or die
harry

[Wodisch]

Hello Richard,

I would not like to spoil your perl-isms, but if your rae talking about "raw" vs. "cooked" mode of the shell input, you could use statements like these:

#!/usr/bin/ksh
key="."
old=$(stty -g)
stty -echo raw
while echo "Enter key by key:"
[ "$key" != "q" ]
do key=$(dd bs=1 count=1 2>/dev/null)
echo "key = $(echo '$key'|cat -v)"
done
stty $old

Not very elegant, ok ;-)
Wodisch