RE: regexp question
Hello all,
Thank you for the time. I guess that this foray into lookaround assertions
was bust. ;0) I am using Parse::RecDescent and I realized that I could use
the
order of Rules to ensure that if I *wasn't* looking at either a \\n or a \
then I would have a bare \ that a simple Production could deal with.
Thanks (Mr $) to all!
chahn
-Original Message-
From: $Bill Luebkert [mailto:[EMAIL PROTECTED]
Sent: Tuesday, December 14, 2004 7:37 PM
To: Christopher Hahn
Cc: [EMAIL PROTECTED]
Subject: Re: regexp question
Christopher Hahn wrote:
$Bill,
The (?: ) construct may be non-capturing, but it does eat text from
the buffer (sic)
...and, besides, when I ran it I saw this:
=
1: asd asdf adf asd \n asd adf
2: asd asdf adf asd \n asd adf
3: asd asdf adf asd asd adf
=
where I need to see something like this:
=
1: asd asdf adf asd
2: asd asdf adf asd \n asd adf
3: asd asdf adf asd asd adf
4: asd asdf adf asd asd adf ad dasf \n dsaf
=
i.e. \n should pass through, where \\n or \ should not.
Does a non-greedy match help (also - I was missing an escape or two on the
\\n):
foreach (
'1: asd asdf adf asd n asd adf \ ad dasf dsaf ',
'2: asd asdf adf asd \n asd adf \ ad dasf dsaf ',
'3: asd asdf adf asd asd adf \ ad dasf n dsaf ',
'4: asd asdf adf asd asd adf ad dasf \n dsaf ',
) {
if (/^(.*?)(?:n|\\)/) {
print $1\n;
}
}
What about trying something like:
$strval =~ m/^(( [^\\] | \\ (?! \\ (?= n)) | \\ (?! \) )*)/x;
which (or so I think ;0) collects as many characters from the
beginning of $str that meet these conditions (bs = backslash):
=
1) not a bs
or 2) a bs that is *not* followed by a ( bs that *is* followed by a n
) or 3) a bs that is *not* followed by a
=
--
,-/- __ _ _ $Bill LuebkertMailto:[EMAIL PROTECTED]
(_/ / )// // DBE CollectiblesMailto:[EMAIL PROTECTED]
/ ) /-- o // // Castle of Medieval Myth Magic
http://www.todbe.com/
-/-' /___/__/_/_http://dbecoll.tripod.com/ (My Perl/Lakers stuff)
___
Perl-Win32-Users mailing list
[EMAIL PROTECTED]
To unsubscribe: http://listserv.ActiveState.com/mailman/mysubs
Re: regexp question
I seem to be missing a piece of the puzzleI want to define a character class ([]) with atoms to (not) match involve negative look-ahead assertions.but no joy. I am reading a stream of text that may contain the two segments \\n and \" I want to define a regexp that will match up to the first of either of these.
I would suggest a slightly different approach, especially since you apparently might not have either sequence. Try
if ($str =~ /n|\\"/)
{
$mytext = substr($str, 0, $-[0]);
# do whatever with $mytext
}
The arrays @- and @+ contain the starting and ending offsets within a string of the portion successfully matched by a regular _expression_; the 0th elements refer to the whole regex, and other elements to the portions captured by parentheses.
See perlvar in the docs and the description of substr under perlfunc.
--Eric AmickColumbia, MD
___
Perl-Win32-Users mailing list
[EMAIL PROTECTED]
To unsubscribe: http://listserv.ActiveState.com/mailman/mysubs
regexp question
Hey, I seem to be missing a piece of the puzzleI want to define a character class ([]) with atoms to (not) match involve negative look-ahead assertions.but no joy. I am reading a stream of text that may contain the two segments \\n and \ I want to define a regexp that will match up to the first of either of these. ...ie. something like ([^]*) where the character class is just the two sequences above. ...but they are not characters at all, but strings, and so I wonder how to approach this. Question: how best to do something to set $1 == every character in the string up to and not including the first of either a \\n or a \ That is all. (something like $strval =~ m/ (.* (?! \\ (?= \ | \\ (?= n) ) ) )/x;) I am going to use this regexp in a Parse::RecDescent Production, and have other Rules to deal with the \\n and \ strings. I am banging on this and will report when something good comes out of it, but please do chime in with any best practices that suggest themselves to you. TIA! Christopher -- Realisant mon espoir, je me lance vers la gloire Christopher Kenneth Hahn -- [EMAIL PROTECTED] ___ Perl-Win32-Users mailing list [EMAIL PROTECTED] To unsubscribe: http://listserv.ActiveState.com/mailman/mysubs
Re: regexp question
Christopher Hahn wrote:
Hey,
I seem to be missing a piece of the puzzleI want to define a character
class ([])
with atoms to (not) match involve negative look-ahead assertions.but no
joy.
I am reading a stream of text that may contain the two segments \\n and \
I want to define a regexp that will match up to the first of either of
these.
...ie. something like ([^]*) where the character class is just the two
sequences above.
...but they are not characters at all, but strings, and so I wonder how to
approach this.
Question: how best to do something to set
$1 == every character in the string up to and not including the first of
either a \\n or a \
That is all. (something like $strval =~ m/ (.* (?! \\ (?= \ | \\ (?= n) ) )
)/x;)
I am going to use this regexp in a Parse::RecDescent Production, and have
other Rules to
deal with the \\n and \ strings.
I am banging on this and will report when something good comes out of it,
but please do
chime in with any best practices that suggest themselves to you.
I assume that \\n is actually 3 characters and not a newline.
It should be as simple as :
use strict;
foreach (
'1: asd asdf adf asd \\n asd adf \ ad dasf dsaf ',
'2: asd asdf adf asd \n asd adf \ ad dasf dsaf ',
'3: asd asdf adf asd asd adf \ ad dasf \\n dsaf ',
'4: asd asdf adf asd asd adf ad dasf \n dsaf ',
) {
if (/^(.*)(?:n|\\)/) {
print $1\n;
}
}
__END__
--
,-/- __ _ _ $Bill LuebkertMailto:[EMAIL PROTECTED]
(_/ / )// // DBE CollectiblesMailto:[EMAIL PROTECTED]
/ ) /-- o // // Castle of Medieval Myth Magic http://www.todbe.com/
-/-' /___/__/_/_http://dbecoll.tripod.com/ (My Perl/Lakers stuff)
___
Perl-Win32-Users mailing list
[EMAIL PROTECTED]
To unsubscribe: http://listserv.ActiveState.com/mailman/mysubs
RE: regexp question
$Bill,
The (?: ) construct may be non-capturing, but it does eat text from the
buffer (sic)
...and, besides, when I ran it I saw this:
=
1: asd asdf adf asd \n asd adf
2: asd asdf adf asd \n asd adf
3: asd asdf adf asd asd adf
=
where I need to see something like this:
=
1: asd asdf adf asd
2: asd asdf adf asd \n asd adf
3: asd asdf adf asd asd adf
4: asd asdf adf asd asd adf ad dasf \n dsaf
=
i.e. \n should pass through, where \\n or \ should not.
What about trying something like:
$strval =~ m/^(( [^\\] | \\ (?! \\ (?= n)) | \\ (?! \) )*)/x;
which (or so I think ;0) collects as many characters from the beginning of
$str that
meet these conditions (bs = backslash):
=
1) not a bs
or 2) a bs that is *not* followed by a ( bs that *is* followed by a n )
or 3) a bs that is *not* followed by a
=
Am I making any sense? ;0)
Thank you for taking the time in any case!
Christopher
-Original Message-
From: $Bill Luebkert [mailto:[EMAIL PROTECTED]
Sent: Tuesday, December 14, 2004 1:41 PM
To: Christopher Hahn
Cc: [EMAIL PROTECTED]
Subject: Re: regexp question
Christopher Hahn wrote:
Hey,
I seem to be missing a piece of the puzzleI want to define a
character class ([]) with atoms to (not) match involve negative
look-ahead assertions.but no joy.
I am reading a stream of text that may contain the two segments \\n and \
I want to define a regexp that will match up to the first of either of
these.
...ie. something like ([^]*) where the character class is just the two
sequences above.
...but they are not characters at all, but strings, and so I wonder
how to approach this.
Question: how best to do something to set
$1 == every character in the string up to and not including the
first of either a \\n or a \
That is all. (something like $strval =~ m/ (.* (?! \\ (?= \ | \\ (?=
n) ) )
)/x;)
I am going to use this regexp in a Parse::RecDescent Production, and
have other Rules to deal with the \\n and \ strings.
I am banging on this and will report when something good comes out of
it, but please do chime in with any best practices that suggest
themselves to you.
I assume that \\n is actually 3 characters and not a newline.
It should be as simple as :
use strict;
foreach (
'1: asd asdf adf asd \\n asd adf \ ad dasf dsaf ',
'2: asd asdf adf asd \n asd adf \ ad dasf dsaf ',
'3: asd asdf adf asd asd adf \ ad dasf \\n dsaf ',
'4: asd asdf adf asd asd adf ad dasf \n dsaf ',
) {
if (/^(.*)(?:n|\\)/) {
print $1\n;
}
}
__END__
--
,-/- __ _ _ $Bill LuebkertMailto:[EMAIL PROTECTED]
(_/ / )// // DBE CollectiblesMailto:[EMAIL PROTECTED]
/ ) /-- o // // Castle of Medieval Myth Magic
http://www.todbe.com/
-/-' /___/__/_/_http://dbecoll.tripod.com/ (My Perl/Lakers stuff)
___
Perl-Win32-Users mailing list
[EMAIL PROTECTED]
To unsubscribe: http://listserv.ActiveState.com/mailman/mysubs
Re: regexp question
Christopher Hahn wrote:
$Bill,
The (?: ) construct may be non-capturing, but it does eat text from the
buffer (sic)
...and, besides, when I ran it I saw this:
=
1: asd asdf adf asd \n asd adf
2: asd asdf adf asd \n asd adf
3: asd asdf adf asd asd adf
=
where I need to see something like this:
=
1: asd asdf adf asd
2: asd asdf adf asd \n asd adf
3: asd asdf adf asd asd adf
4: asd asdf adf asd asd adf ad dasf \n dsaf
=
i.e. \n should pass through, where \\n or \ should not.
Does a non-greedy match help (also - I was missing an escape
or two on the \\n):
foreach (
'1: asd asdf adf asd n asd adf \ ad dasf dsaf ',
'2: asd asdf adf asd \n asd adf \ ad dasf dsaf ',
'3: asd asdf adf asd asd adf \ ad dasf n dsaf ',
'4: asd asdf adf asd asd adf ad dasf \n dsaf ',
) {
if (/^(.*?)(?:n|\\)/) {
print $1\n;
}
}
What about trying something like:
$strval =~ m/^(( [^\\] | \\ (?! \\ (?= n)) | \\ (?! \) )*)/x;
which (or so I think ;0) collects as many characters from the beginning of
$str that
meet these conditions (bs = backslash):
=
1) not a bs
or 2) a bs that is *not* followed by a ( bs that *is* followed by a n )
or 3) a bs that is *not* followed by a
=
--
,-/- __ _ _ $Bill LuebkertMailto:[EMAIL PROTECTED]
(_/ / )// // DBE CollectiblesMailto:[EMAIL PROTECTED]
/ ) /-- o // // Castle of Medieval Myth Magic http://www.todbe.com/
-/-' /___/__/_/_http://dbecoll.tripod.com/ (My Perl/Lakers stuff)
___
Perl-Win32-Users mailing list
[EMAIL PROTECTED]
To unsubscribe: http://listserv.ActiveState.com/mailman/mysubs
Re: RegExp Question
John Giordano wrote:
$grep_deferred = system ('findstr DeferredStatus response1');
print "$grep_deferred\n\n";
[snip]
$grep_deferred has this in it:
a href="/c9410ee04de9845704db8951dfde015b/DeferredStatus"img
src="/images/btnstats.gif" width=120 height=40 border=0 alt="Mail
Status"/a
Are you sure? Try printing out "bloop\nblip\n$grep_deferred\nblep\n" and see
whether this a href line comes between "blip" and "blep".
From `perldoc -f system`:
The return value is the exit status of the program as returned
by the `wait' call. To get the actual exit value divide by 256.
So if $grep_deferred is the return value from system, it's probably a
number. The line you're seeing on the screen is presumably the output of
findstr, which went to STDOUT; since you didn't redirect STDOUT, it went the
same place your print went.
You probably want `` (backticks).
ObTMTOWTDI: if you're looking for a fixed substring such as 'a ', then
consider using index() rather than regular expressions. See useless
benchmark at the end.
Cheers,
Philip
#!perl -w
use strict;
use Benchmark 'cmpthese';
cmpthese(5_000_000, {
'index.start.found'= sub { $a = index 'a
href="/c9410ee04de9845704db8951dfde015b/DeferredStatus"img
src="/images/btnstats.gif" width=120 height=40 border=0 alt="Mail
Status"/a', 'a '; },
'index.start.notfound' = sub { $a = index '[a
href="/c9410ee04de9845704db8951dfde015b/DeferredStatus"img
src="/images/btnstats.gif" width=120 height=40 border=0 alt="Mail
Status"/a', 'a '; },
'regex.start.found'= sub { $a = 'a
href="/c9410ee04de9845704db8951dfde015b/DeferredStatus"img
src="/images/btnstats.gif" width=120 height=40 border=0 alt="Mail
Status"/a' =~ /a /; },
'regex.start.notfound' = sub { $a = '[a
href="/c9410ee04de9845704db8951dfde015b/DeferredStatus"img
src="/images/btnstats.gif" width=120 height=40 border=0 alt="Mail
Status"/a' =~ /a /; },
'index.end.found' = sub { $a = index 'fhuhge8a goija fgoja w04gua
roigj oöijf g9a8u onigö lfsdkj gölija osij ga0 jügojar öoigj öosijd ü08g9u
gaji fäöigjah+w r0g9j aäüjäfpoiajs öoijfd +ajsd fäüpioaj öwoeijf öoasuidhf
p9iawh üefhaj woeijg awopihg apiuehrg üaiowj eoöfigjaw ejifa
href="/c9410ee04de9845704db8951dfde015b/DeferredStatus"img
src="/images/btnstats.gif" width=120 height=40 border=0 alt="Mail
Status"/a', 'a '; },
'index.end.notfound' = sub { $a = index 'fhuhge8a goija fgoja w04gua
roigj oöijf g9a8u onigö lfsdkj gölija osij ga0 jügojar öoigj öosijd ü08g9u
gaji fäöigjah+w r0g9j aäüjäfpoiajs öoijfd +ajsd fäüpioaj öwoeijf öoasuidhf
p9iawh üefhaj woeijg awopihg apiuehrg üaiowj eoöfigjaw ejif[a
href="/c9410ee04de9845704db8951dfde015b/DeferredStatus"img
src="/images/btnstats.gif" width=120 height=40 border=0 alt="Mail
Status"/a', 'a '; },
'regex.end.found' = sub { $a = 'fhuhge8a goija fgoja w04gua roigj
oöijf g9a8u onigö lfsdkj gölija osij ga0 jügojar öoigj öosijd ü08g9u gaji
fäöigjah+w r0g9j aäüjäfpoiajs öoijfd +ajsd fäüpioaj öwoeijf öoasuidhf p9iawh
üefhaj woeijg awopihg apiuehrg üaiowj eoöfigjaw ejifa
href="/c9410ee04de9845704db8951dfde015b/DeferredStatus"img
src="/images/btnstats.gif" width=120 height=40 border=0 alt="Mail
Status"/a' =~ /a /; },
'regex.end.notfound' = sub { $a = 'fhuhge8a goija fgoja w04gua roigj
oöijf g9a8u onigö lfsdkj gölija osij ga0 jügojar öoigj öosijd ü08g9u gaji
fäöigjah+w r0g9j aäüjäfpoiajs öoijfd +ajsd fäüpioaj öwoeijf öoasuidhf p9iawh
üefhaj woeijg awopihg apiuehrg üaiowj eoöfigjaw ejif[a
href="/c9410ee04de9845704db8951dfde015b/DeferredStatus"img
src="/images/btnstats.gif" width=120 height=40 border=0 alt="Mail
Status"/a' =~ /a /; },
'assign' = sub { $a = 1; },
});
__END__
Output:
Benchmark: timing 500 iterations of assign, index.end.found,
index.end.notfound, index.start.found, inde
x.start.notfound, regex.end.found, regex.end.notfound, regex.start.found,
regex.start.notfound...
assign: 0 wallclock secs ( 1.02 usr + 0.00 sys = 1.02 CPU) @
4897159.65/s (n=500)
index.end.found: 9 wallclock secs ( 9.79 usr + 0.00 sys = 9.79 CPU) @
510516.64/s (n=500)
index.end.notfound: 12 wallclock secs (12.38 usr + 0.00 sys = 12.38 CPU) @
404007.76/s (n=500)
index.start.found: 3 wallclock secs ( 2.49 usr + 0.00 sys = 2.49 CPU) @
2004811.55/s (n=500)
index.start.notfound: 5 wallclock secs ( 5.42 usr + 0.00 sys = 5.42 CPU)
@ 922849.76/s (n=500)
regex.end.found: 12 wallclock secs (11.42 usr + 0.00 sys = 11.42 CPU) @
437943.42/s (n=500)
regex.end.notfound: 13 wallclock secs (13.34 usr + 0.00 sys = 13.34 CPU) @
374840.69/s (n=500)
regex.start.found: 5 wallclock secs ( 3.91 usr + 0.00 sys = 3.91 CPU) @
1280081.93/s (n=500)
regex.start.notfound: 7 wallclock secs ( 6.38 usr + 0.00 sys = 6.38 CPU)
@ 783821.92/s (n=500)
Rate regex.end.notfound index.end.notfound
regex.end.found index.end.found regex.start.notfound index.start.notfound
regex.start.found index.start.found assign
regex.end.notfound374841/s
