RegEx Help, Please? (match after n)

[Smarta55 Chris]

Can someone help me find the last n folders in a path,
or the last n OUs in a fully qualified Netware name... for an app object,
etc?

Example:
\\server\volume\share\folder\folder\folder
I want to grab just the last 2 folders (including or excluding the \s), but
sometimes there
are 3 folders after the share, like in the example, and sometimes there are
more.

Same for an app object in eDirectory
..appobject.dept.biz.site.county.state.tree
How to I grab just the .site.county.state.tree OUs?
Sometimes there are 5 OUs after the app object, and sometimes more.
I want, I think, to "start at the end of the string, and grab everything
after the 4th dot from the end," but don't have a clue how to code this.

I'm not a perl newbie, but I'm no expert....and I AM a RegEx idiot!

Here's some of what I've tried:
$string = ".appobject.dept.biz.site.county.state.tree";
$string =~ /\..+\..+\..+/$; # i.e. match a literal dot followed my
anything mult times, followed by another literal dot, followed, etc...
# starting at the end of the string
$string =~ /(\.\W)*/; # \w does NOT match AlphaNum like it's supposed to,
\W DOES, but it's NOT supposed to.
# unless . is an AlphaNum character, but I'm
not sure

And I can't remember the 4,612 other patterns I've tried, or the
accompanying 4,612 hairs I've
yanked from my scalp in the process. :-(

Thank you,
Chris

 

[Gunnar Hjalmarsson]

Can someone help me find the last n folders in a path,
or the last n OUs in a fully qualified Netware name... for an app object,
etc?

Example:
\\server\volume\share\folder\folder\folder
I want to grab just the last 2 folders (including or excluding the \s), but
sometimes there
are 3 folders after the share, like in the example, and sometimes there are
more.

Same for an app object in eDirectory
.appobject.dept.biz.site.county.state.tree
How to I grab just the .site.county.state.tree OUs?
Sometimes there are 5 OUs after the app object, and sometimes more.
I want, I think, to "start at the end of the string, and grab everything
after the 4th dot from the end," but don't have a clue how to code this.

I'm not a perl newbie, but I'm no expert....and I AM a RegEx idiot!

Then don't overuse them. How about this for your first example:

my @folders = ( split /\\/, $path )[-2..-1];

Leaving the second example as an exercise for you. ;-)

 

[Smarta55 Chris]

Thank you. I used:
my @ous = (split /\.,/,$string)[-2..-1]; and got back 'folder folder'.
changing
-2 to -3 brought back 'folder folder folder'
I've used split before, but I don't understand the [-2..-1] part.
I ruled out using split because I'll never know how long the string
will be, or how many OUs will follow the object in its FQN.
Also, it's not returning what I need....obviously my thought/need wasn't
translated well to my text..

When I match against ".appobject.dept.biz.site.county.state.tree", and
need to return the last 4 OUs, what I need back is ".site.county.state.tree"
The leading dot before
site isn't important (but is preferred), but the 3 dots between site,
county, state, and tree are needed.
I plan to append a new object, and additional OU(s), before what's returned.

Thank you for your help,
Chris

 

[A. Sinan Unur]

Can someone help me find the last n folders in a path,
or the last n OUs

You are missing an 'I' there ... Seriously, though, I don't know what
you are referring to.

Example:
\\server\volume\share\folder\folder\folder
I want to grab just the last 2 folders (including or excluding the
\s), but sometimes there are 3 folders after the share, like in the
example, and sometimes there are more.
OK.

Same for an app object in eDirectory
.appobject.dept.biz.site.county.state.tree
How to I grab just the .site.county.state.tree OUs?
Sometimes there are 5 OUs after the app object, and sometimes more.
I want, I think, to "start at the end of the string, and grab
everything after the 4th dot from the end," but don't have a clue how
to code this.

perldoc perlfunc
perldoc -f reverse
perldoc -f split

In addition,

perldoc -f rindex
perldoc -f substr

might also be of interest, in case you want to write a different
algorithm.

I'm not a perl newbie, but I'm no expert....and I AM a RegEx idiot!

Probably not, but you need to look at this calmly.

Here's some of what I've tried:
$string = ".appobject.dept.biz.site.county.state.tree";
$string =~ /\..+\..+\..+/$; # i.e. match a literal dot followed my
anything mult times, followed by another literal dot, followed, etc...
# starting at the end of the
string
$string =~ /(\.\W)*/; # \w does NOT match AlphaNum like it's
supposed to, \W DOES, but it's NOT supposed to.
# unless . is an AlphaNum character,
but I'm
not sure

This looks like Def Poetry to me.

And I can't remember the 4,612 other patterns I've tried, or the
accompanying 4,612 hairs I've yanked from my scalp in the process.
:-(

Here are two examples to get you started:

First:

#!/usr/bin/perl

use strict;
use warnings;

sub last_n_components {
my ($s, $d, $n) = @_;
join($d, reverse((reverse split /\Q$d\E/, $s) [0 .. $n - 1]))
}

print +last_n_components(
q{.appobject.dept.biz.site.county.state.tree},
'.', 2
)."\n";

print +last_n_components(
q{\\\\server\volume\share\folder\folder\folder},
'\\', 2
)."\n";

__END__

Now, if you are going to deal with paths, your life will be much easier
if you do it portably from the start:

#!/usr/bin/perl

use strict;
use warnings;

use File::Spec::Functions qw'catdir splitdir';

sub last_n_path_components {
my ($s, $n) = @_;
catdir reverse((reverse splitdir $s)[0 .. $n - 1]);
}

print +last_n_path_components(
q{\\\\server\volume\share\folder\folder\folder}, 2)."\n";

__END__

 

[A. Sinan Unur]

"Smarta55 Chris" <smarta55ATcomcastDOTnet> wrote in

[ Top-posting fixed. Please do not do that. ]

Then don't overuse them. How about this for your first example:

my @folders = ( split /\\/, $path )[-2..-1];

Leaving the second example as an exercise for you. ;-)

my @ous = (split /\.,/,$string)[-2..-1]; and got back 'folder
folder'.
changing
-2 to -3 brought back 'folder folder folder'

You mentioned something about not being a newbie. You do need to read
the first few chapters of 'Learning Perl'.

@ous is an array. You can join the elements of that array to get what
ever you want.

I've used split before, but I don't understand the [-2..-1] part.

That is an array slice. It selects all the elements from the second to
last to the last one from the array returned by split. It is infinitely
more efficient than the double reverse I used.

I ruled out using split because I'll never know how long the string
will be, or how many OUs will follow the object in its FQN.

WTF? Is this an acronym contest?

Anyway, by indexing from the end of the array, you do not need to know
how many components in total there are, because, as you said, you are
only interested in the last n.

Also, it's not returning what I need ...

Yes it is. You just do not know Perl very well.

When I match against ".appobject.dept.biz.site.county.state.tree",
and need to return the last 4 OUs, what I need back is
".site.county.state.tree" The leading dot before
site isn't important (but is preferred), but the 3 dots between site,
county, state, and tree are needed.

Then join the elements of the returned array using a dot.

That is what

perldoc -f join

is for.

Sinan

 

[A. Sinan Unur]

sub last_n_components {
my ($s, $d, $n) = @_;
join($d, reverse((reverse split /\Q$d\E/, $s) [0 .. $n - 1]))
}

Ahem ... This is what happens if you try to watch TV and post at the
same time.

Sorry about that.

This should be better:

#!/usr/bin/perl

use strict;
use warnings;

use File::Spec::Functions qw'catdir splitdir';

sub last_n_components {
my ($s, $d, $n) = @_;
join $d, (split /\Q$d\E/, $s)[-$n .. -1];
}

sub last_n_path_components {
my ($s, $n) = @_;
catdir( (splitdir $s)[-$n .. - 1] );
}

print +last_n_components(
q{.appobject.dept.biz.site.county.state.tree},
'.', 2
)."\n";

print +last_n_path_components(
q{\\\\server\volume\share\folder\folder\folder}, 2)."\n";

__END__

 

[Smarta55 Chris]

Thank you....for teaching me that I am, afterall, a perl newbie.
I'm more confused than before. I'll check the perldocs you
mention, then assign this script to someone else at work. :-(

There's no simple, single-line $ao =~ /some pattern or other/; that'll work?
I just want to append an app object's name to a biz name, then to the last 4
OUs of the match, and create
it in eDirectory with a 3rd party utility. :-(

Thanks for your help!
Chris

Can someone help me find the last n folders in a path,
or the last n OUs

You are missing an 'I' there ... Seriously, though, I don't know what
you are referring to.

Example:
\\server\volume\share\folder\folder\folder
I want to grab just the last 2 folders (including or excluding the
\s), but sometimes there are 3 folders after the share, like in the
example, and sometimes there are more.
OK.

Same for an app object in eDirectory
.appobject.dept.biz.site.county.state.tree
How to I grab just the .site.county.state.tree OUs?
Sometimes there are 5 OUs after the app object, and sometimes more.
I want, I think, to "start at the end of the string, and grab
everything after the 4th dot from the end," but don't have a clue how
to code this.

perldoc perlfunc
perldoc -f reverse
perldoc -f split

In addition,

perldoc -f rindex
perldoc -f substr

might also be of interest, in case you want to write a different
algorithm.

I'm not a perl newbie, but I'm no expert....and I AM a RegEx idiot!

Probably not, but you need to look at this calmly.

Here's some of what I've tried:
$string = ".appobject.dept.biz.site.county.state.tree";
$string =~ /\..+\..+\..+/$; # i.e. match a literal dot followed my
anything mult times, followed by another literal dot, followed, etc...
# starting at the end of the
string
$string =~ /(\.\W)*/; # \w does NOT match AlphaNum like it's
supposed to, \W DOES, but it's NOT supposed to.
# unless . is an AlphaNum character,
but I'm
not sure

This looks like Def Poetry to me.

And I can't remember the 4,612 other patterns I've tried, or the
accompanying 4,612 hairs I've yanked from my scalp in the process.
:-(

Here are two examples to get you started:

First:

#!/usr/bin/perl

use strict;
use warnings;

sub last_n_components {
my ($s, $d, $n) = @_;
join($d, reverse((reverse split /\Q$d\E/, $s) [0 .. $n - 1]))
}

print +last_n_components(
q{.appobject.dept.biz.site.county.state.tree},
'.', 2
)."\n";

print +last_n_components(
q{\\\\server\volume\share\folder\folder\folder},
'\\', 2
)."\n";

__END__

Now, if you are going to deal with paths, your life will be much easier
if you do it portably from the start:

#!/usr/bin/perl

use strict;
use warnings;

use File::Spec::Functions qw'catdir splitdir';

sub last_n_path_components {
my ($s, $n) = @_;
catdir reverse((reverse splitdir $s)[0 .. $n - 1]);
}

print +last_n_path_components(
q{\\\\server\volume\share\folder\folder\folder}, 2)."\n";

__END__

 

[A. Sinan Unur]

Thank you....

You are welcome. However, please note that top-posting and full-quoting
are generally not useful in facilitating a productive exchange.

Please do read the posting guidelines for this group. They contain
valuable information on how you can help yourself, and help others help
you.

for teaching me that I am, afterall, a perl newbie.

I'm more confused than before. I'll check the perldocs you
mention, then assign this script to someone else at work. :-(

Well, to be honest, my double reverse was very convoluted, and I have
onlyself (no, actually, HBO) to blame.

There's no simple, single-line $ao =~ /some pattern or other/;

There is no reason to bring out the big guns for this kind of thing.
Besides, I would probably mess that up even worse than I messed up with
the double reverse.

Sinan

 

[Smarta55 Chris]

You are welcome. However, please note that top-posting and full-quoting
are generally not useful in facilitating a productive exchange.

Please do read the posting guidelines for this group. They contain
valuable information on how you can help yourself, and help others help
you.

Now I'm a newsgroups newbie, too!

Thanks

 

[Gunnar Hjalmarsson]

Thank you. I used:
my @ous = (split /\.,/,$string)[-2..-1]; and got back 'folder folder'. --------------------------^
??

changing
-2 to -3 brought back 'folder folder folder'
I've used split before, but I don't understand the [-2..-1] part.

split() returns a list, and the above picks a slice of the list.

-1 represents the last, and -2 represents the second last element of a
list, so -2..-1 are 'all' (two) elements between the second last and the
last. See "perldoc perldata".

Sinan made use of the reverse() function instead in his solutions.

I ruled out using split because I'll never know how long the string
will be, or how many OUs will follow the object in its FQN.
Also, it's not returning what I need....obviously my thought/need wasn't
translated well to my text..

When I match against ".appobject.dept.biz.site.county.state.tree", and
need to return the last 4 OUs, what I need back is ".site.county.state.tree"
The leading dot before
site isn't important (but is preferred), but the 3 dots between site,
county, state, and tree are needed.

Well, you could use join(). Or maybe just:

my $n = 4;
print $& if $string =~ /(?:\.[^.]+){$n}$/;

 

[A. Sinan Unur]

Now I'm a newsgroups newbie, too!

Well, we all started somewhere

Now, the next thing to take care of is to make sure you properly
attribute the quoted portions, so that people can keep track of who said
what.

I see you are using Outlook Express on Windows: May I recommend XNews as
a proper newsreader?

<URL:http://xnews.3dnews.net/>

Sinan

 

[Gunnar Hjalmarsson]

I've used split before, but I don't understand the [-2..-1] part.

That is an array slice. It selects all the elements from the second to
last to the last one from the array returned by split.

s/array/list/g;

 

[Dave A.]

There's no simple, single-line $ao =~ /some pattern or other/; that'll work?

#!perl -w

my $n = 4; # number of folders to match
my $regexp = qr/((?:\\[^\\]+){$n})$/;

while (<STDIN>) {
chomp;
(my $folders) = m/$regexp/; # matches only if 4 or more folders
print "$folders\n";

# or
# /((?:\\[^\\]+){MIN,MAX})$/
# match *between* MIN and MAX number of folders

# or
# /((?:\\[^\\]+){MIN,})$/
# match *at least* MIN number of folders

# etc...

}
__END__

Dave

 

[Dave A.]

Smarta55 Chris wrote:

I want to stress that I would never use a regular expression in a
situation such as this, where a simple *split* is clearly the right tool
for the job. The code below was intended as an example only.

Why use error-prone regular expressions when one can write, e.g.,
@folders = split( m![\\/]!, $path )

Dave

#!perl -w

my $n = 4; # number of folders to match
my $regexp = qr/((?:\\[^\\]+){$n})$/;

while (<STDIN>) {
chomp;
(my $folders) = m/$regexp/; # matches only if 4 or more folders
print "$folders\n";

# or
# /((?:\\[^\\]+){MIN,MAX})$/
# match *between* MIN and MAX number of folders

# or
# /((?:\\[^\\]+){MIN,})$/
# match *at least* MIN number of folders

# etc...

}
__END__