|
|
 Up |
|
|
  |
Author: JackJack
Date: Jul 14, 2006 23:49
Hi folks,
Ok I am stumped... I am reading a fileline in as follows:
12535957@140@2006-07-08 09:34:45.000@@@2005-11-24@
and splitting on the "@"...
open(SOURCE1,$filename1)
@columnarray = split(/\@/, $_);
Whats really wierd is $columnarray[4] has a length of 1, when clearly
above its NULL !!
I want to be able to recognize this as null, but you cant even match
regex it... I tried
if ($columnarray[4] =~ m/.*?[[:print:]]/) { print " NON PRINTABLE
"."\n"; }
$columnarray[4] =~ s/^\s+//;
$columnarray[4] =~ s/\s+$//;
$length = length @columnarray[4]."\n";
and a number of other things that dont work.. how do I identify "WHAT"
is making up the length =1 ?
Also, what is the detection match test for NULL ?
|
| Show full article (0.75Kb) |
|
| |
3 Comments |
|
|
Author: Ben MorrowBen Morrow
Date: Jul 14, 2006 23:25
Quoth "A. Sinan Unur" <1usa@llenroc.ude.invalid>:
>>
>> Quoth "scottmf" gmail.com>:>>>> I'm fairly sure you still need the Win32::OLE->new line. If matlab
>>>> is running it should give you a handle on the current instance.
>>>> [Caveat: my understanding of OLE is sketchy at best :)]
>>>>
>>>> The Perl program is in no sense 'inside' matlab. It is a completely
>>>> separate process: it doesn't even know you invoked it from matlab,
>>>> so how would you get an OLE object?
>>>
>>> when I use the Win32::OLE->new line it launches a new instance of
>>> matlab rather than linking to the current instance. I think this is
>>> the basic problem; that I need some way to link the perl script to
>>> the current matlab instance.
>>
>> OK, you are reading the Win32::OLE docs, yes? About 30 seconds worth ...
|
| Show full article (3.65Kb) |
|
| |
no comments
|
|
|
Author:
Date: Jul 14, 2006 23:04
On Fri, 14 Jul 2006 20:05:33 GMT, A. Sinan Unur <1usa@llenroc.ude.invalid> wrote:
> andreas a@t mrs d.t ch wrote in
> news:44b7f766$0$12613$5402220f@news.sunrise.ch:
>
>> perl -e 'foreach ("abc" =~ m/(.)/gc) {print "$1"}'
>>
>> prints 'ccc' and not what I (and maybe other stupid users) would
>> expect 'abc'.
>
> "abc" =~ m/(.)/gc
>
> in list context returns all the matches. However, you are not printing
> elements of that list, only $1.
> On the other hand, a while loop is better here: It allows you iterate
> through matches without a list of all matches being constructed. That
> can become a significant overhead if the number of matches is large
> (just like the difference between
Ok, I guess I stay with the while, it's more the way I think.
|
| Show full article (0.95Kb) |
|
no comments
|
|
|
Author:
Date: Jul 14, 2006 22:58
On Fri, 14 Jul 2006 23:01:22 +0200, Dr.Ruud wrote:
> andreas a@t mrs d.t ch schreef:
>
>> perl -e 'foreach ("abc" =~ m/(.)/gc) {print "$1"}'
>>
>> prints 'ccc'
>
> Yes. What puzzles me is the c-modifier, why did you use it?
>
Puzzles me to, it somehow sneaked in during my shell examples. The
original regex doesn't use it.
Andreas
|
| |
|
no comments
|
|
|
Author:
Date: Jul 14, 2006 22:56
On 14 Jul 2006 13:09:07 -0700, Paul Lalli gmail.com> wrote:> andreas a@t mrs d.t ch wrote:
>> The problem:
>>
>> perl -e 'foreach ("abc" =~ m/(.)/gc) {print "$1"}'
>>
>> prints 'ccc' and not what I (and maybe other stupid users) would expect
>> 'abc'.
>
> You have several Perl features conspiring against you.
Well... looks like the big conspiration, lets start a theory on that :-)
> 1) the condition to a foreach() loop is evaluated in list context
Took me several minutes to understand this, but I guess I do it now.
The regexp is evaluated correctly, all the matches are stored in a list
and the list is evaluated with foreach, but as the evaluation of the
regexp is done, all the matches are through, $1 is set to the last match
and printed three times for each list element.
|
| Show full article (1.63Kb) |
|
no comments
|
|
|
Author: A. Sinan UnurA. Sinan Unur
Date: Jul 14, 2006 22:00
>
> Quoth "scottmf" gmail.com>:>>> I'm fairly sure you still need the Win32::OLE->new line. If matlab
>>> is running it should give you a handle on the current instance.
>>> [Caveat...
|
| Show full article (2.89Kb) |
|
no comments
|
|
|
Author: Tad McClellanTad McClellan
Date: Jul 14, 2006 20:53
John Bokma castleamber.com> wrote:> Tad McClellan augustmail.com> wrote:
> >> [snip rancid TOFU]
>
> Never liked tofu much. Rancid tofu, there goes my appetite :-(
I think I listened to too much Zappa when I was younger.
--
Tad McClellan SGML consulting
tadmc@ augustmail.com Perl programming
Fort Worth, Texas
|
| |
|
no comments
|
|
|
Author: Mark Jason DominusMark Jason Dominus
Date: Jul 14, 2006 20:32
There is still time to propose a lightning talk for OSCON 2006. The
deadline has been extended through Monday, 25 July.
Lightning Talks at the 2006 O'Reilly Open Source Convention
Lightning talks are brief (5-minute) talks that focus on a single
example, idea, project, or technique. Lightning talks do not attempt
to cover all aspects of their subject matter, but rather to present
one facet of the idea clearly and succinctly.
To submit a proposal for a lightning talk, please send your proposed
title and an abstract of up to four sentences to:
osc-lt-2006-submit@ plover.com
For more complete information, visit:
http://perl.plover.com/lt/osc2006/
Thanks.
|
| |
|
no comments
|
|
|
|
|
|
|
|
|
Author: Tad McClellanTad McClellan
Date: Jul 14, 2006 19:52
> I think you are right about this board.
This is not a "board". This is a Usenet newsgroup.
> I don't think I'm going to
> make it a regular stop on my information super highway.
The loss to this community will be great, but thank you for doing
what you can to make this a better place.
--
Tad McClellan SGML consulting
tadmc@ augustmail.com Perl programming
Fort Worth, Texas
|
| |
|
no comments
|
|
|
|
|
|
|
