Expression problem

[K.J. 44]

I have the two following regular expressions. I am not very good at
writing these yet. I am parsing some logs looking for some key words,
then taking the text after them.

if ($details[$i] =~ /\bworkstation\b\bname:\b\s\b[0-9A-Za-z_\-]+\b/i) {
($nothing, $hostName[$i]) = split(/:/, $&);
}
if ($details[$i] =~ /\buser\b\bname:\b\w+/i) {
print("The username is: $&");
($nothing, $username[$i]) = split(/:/, $&);
}

The $details array is read in from a text file and this works fine.
What I want to do is search the $details text for certain key words,
then take the text right after. The first if statement

Find the word Workstation followed by a space followed by name:
followed by a space followed by a string of characters including word
characters and hyphens. if the match is found, take only the text
after the : as the workstation name.

The second part is along the same lines for username.

Find the word user followed a space followed by the word name: followed
by a space followed by a string of word characters. Split at the : as
the username found.

These do not seem to be finding matches when I can see them in the log
file. Where am I messing up?

Thanks.

 

[Ric]

You need to post at least one of the lines you read from your text file.
You should name the values you would like to have extracted.

example line: yesterday 12:30 hans:went:home

var1 should contain: 12:20
var2 should contain: went

this is much faster and precise than explaining something in a huge
block of text

 

[xhoster]

I have the two following regular expressions. I am not very good at
writing these yet. I am parsing some logs looking for some key words,
then taking the text after them.

if ($details[$i] =~ /\bworkstation\b\bname:\b\s\b[0-9A-Za-z_\-]+\b/i) {

About the only places that \b should be used are the beginning or end
of the regex or before or after something like ".*". Also, I don't see how
you can ever productively have more than one in a row, doing so is probably
exactly the same as having just one.

\b is a zero-width condition. So "tion\b\bname" can never match because
you are demanding that there is a n before the \b zero-width placeholder,
and an n after the \b zero-width placeholder, and if that is the case then
the conditions which define \b are not met. Similarly the \b in ":\b\s" is
impossible to ever match, as it has to have a non-word character on each
side which is what \b does not do. The \b in "\s\b[0-9A-Za-z_\-]" is
almost redundant--it just forbid the "-" from the character class from
being used, which I doubt is what you want.

....

Find the word Workstation followed by a space followed by name:
followed by a space followed by a string of characters including word
characters and hyphens. if the match is found, take only the text
after the : as the workstation name.

.... /\bworkstation name: ([-\w]+)/i ...

If you "space" you meant "white space", then turn my spaces into "\s"
(not "\b").


Xho

 

[K.J. 44]

Thank you very much for your help and suggestions. I will try these
out.

Thanks!

I have the two following regular expressions. I am not very good at
writing these yet. I am parsing some logs looking for some key words,
then taking the text after them.

if ($details[$i] =~ /\bworkstation\b\bname:\b\s\b[0-9A-Za-z_\-]+\b/i) {

About the only places that \b should be used are the beginning or end
of the regex or before or after something like ".*". Also, I don't see how
you can ever productively have more than one in a row, doing so is probably
exactly the same as having just one.

\b is a zero-width condition. So "tion\b\bname" can never match because
you are demanding that there is a n before the \b zero-width placeholder,
and an n after the \b zero-width placeholder, and if that is the case then
the conditions which define \b are not met. Similarly the \b in ":\b\s" is
impossible to ever match, as it has to have a non-word character on each
side which is what \b does not do. The \b in "\s\b[0-9A-Za-z_\-]" is
almost redundant--it just forbid the "-" from the character class from
being used, which I doubt is what you want.

...

Find the word Workstation followed by a space followed by name:
followed by a space followed by a string of characters including word
characters and hyphens. if the match is found, take only the text
after the : as the workstation name.

... /\bworkstation name: ([-\w]+)/i ...

If you "space" you meant "white space", then turn my spaces into "\s"
(not "\b").


Xho

 

[Tad McClellan]

So the first piece of code could be cut down to
if( $details[$i] =~ /\bworkstation\sname:\s([0-9A-Za-z_\-]+)/i )
{
$hostName[$i] = $1;
}

There's still room for improvement, like

adding an //x modifier, particularly if you are "not very good"
at grokking regexes:

if ( $details[$i] =~ /\bworkstation
\s
name:
\s
(
[0-9A-Za-z_-]+
)
/ix
)

e.g. removing A-Z from
the character class, as you are giving the /i modifier anyway,
so both upper and lowercase characters will be matched.

and removing the backslash. Hyphen is not meta in a character
class if it is first or last in the class.

 

[Mr P]

Thank you very much for your help and suggestions. I will try these
out.

Thanks!

I have the two following regular expressions. I am not very good at
writing these yet. I am parsing some logs looking for some key words,
then taking the text after them.

if ($details[$i] =~ /\bworkstation\b\bname:\b\s\b[0-9A-Za-z_\-]+\b/i) {

 

[Mr P]

The $details array is read in from a text file and this works fine.
What I want to do is search the $details text for certain key words,
then take the text right after. The first if statement

 

[Dr.Ruud]

Mr P schreef:

PS: [0-9A-Za-z_\-] looks a LOT like \w

But \w matches 91801 characters (codepoints).
http://www.xs4all.nl/~rvtol/perl/unicount.pl