|
(chromatic) Re: How to substitute all the words before a certain word in the middle of a line?
by chromatic (Archbishop) on Jul 12, 2000 at 09:30 UTC
|
Here's a happy little regex:
my $sentence = "This is a nice weather. He is a nice guy.
We have some nice food in our dinner.";
$sentence =~ s:(\w+\s)+(\bnice\b.*?\.):A $2:g;
print "$sentence\n";
That looks for a word followed by a space, repeated any number of times, followed by the word nice (with word boundaries around it) followed by a minimal number of characters before a final period. It replaces the first bit with 'A' and keeps the second around. /g, of course, makes it look for all the matches in the string. | [reply]
[d/l] |
Re: How to substitute all the words before a certain word in the middle of a line?
by Anonymous Monk on Jul 12, 2000 at 17:31 UTC
|
The first one almost does the job, but it's simpler just to discard every single character before the nice IMHO...
my $sentence = "This is a nice weather. He is a nice guy.
We have some nice food in our dinner.";
$sentence =~ s:.*(\bnice\b.*?\.):A $1:g;
print "$sentence\n";
| [reply] |
RE: How to substitute all the words before a certain word in the middle of a line?
by Ovid (Cardinal) on Jul 12, 2000 at 21:20 UTC
|
Better late than never: here's my take on the situation.
$test = "This is a nice weather. He is a nice guy. We have some nice f
+ood in our dinner.";
$test =~ s/(?:.*?)(\bnice\b[^.]*\.\s*)/A $1/g;
print $test;
The (?: allows grouping without backreferencing, thus making the regex a little more efficient. The [^.] after the word "nice" forces it to match up to, but excluding the period. This is much more efficient than the dot star ".*" because of the way Perl's regex engine works. Unfortunately, I could not figure out a clean way of getting rid of the first dot star.
Cheers,
Ovid | [reply]
[d/l] |
Re: How to substitute all the words before a certain word in the middle of a line?
by vkonovalov (Monk) on Jul 12, 2000 at 09:36 UTC
|
s/\b\w+\s+nice\b/A nice/g;
or
$w = 'nice'; $r = 'A';
s/\b\w+(\s+$w)\b/$r$1/g;
should work | [reply]
[d/l]
[select] |
|
|
The first one does not work, at least not here, when I supply the given sentence.
You just replace the last word before 'nice' with 'A'.
| [reply] |
|
|
More careful reading will help in my case :)
The word "all" was lost by me :)
| [reply] |
RE: How to substitute all the words before a certain word in the middle of a line?
by Emphyrio (Beadle) on Jul 12, 2000 at 13:42 UTC
|
This would be a solution, keeping in mind that we assume all '.' characters terminate a line.
$line="This is a nice weather. He is a nice guy. We have some nice food in our dinner.";
foreach (split /\./,$line) {
s/^.*(nice.*)/$1/;
$result.="A $_.";
}
print $result;
| [reply] |
|
|
I doubt that this code will work properly... I think this algorithm is far away of originally desired one.
| [reply] |
Re: How to substitute all the words before a certain word in the middle of a line?
by Ted Nitz (Chaplain) on Jul 12, 2000 at 20:01 UTC
|
I'm assuming that there should have been line breaks in the sample data, so that each sentince was on a separate line. Also, this code will match everything before the first occurance of the trigger word (nice in this case). You can replace <DATA> with another filehandle or with <> depending upon where you want to get the lines.
my $triggerword = "nice";
my $replacement = "A ";
while(<DATA>){
s/^.*?(?=$triggerword)/$replacement/o;
print;
}
__DATA__
This is a nice weather.
He is a nice guy.
We have some nice food in our dinner.
| [reply]
[d/l]
[select] |
Re: How to substitute all the words before a certain word in the middle of a line?
by ZZamboni (Curate) on Jul 12, 2000 at 22:54 UTC
|
Just because TIMTOWTDI, here's another way. If you know that all your sentences are separated by periods:
foreach (@x=split(/(\.\s+)/, $text)) {
s/^.*(\bnice\b.*)$/A $1/;
}
$text=join("", @x);
Don't go with it, though. chromatic's regex is much nicer :-)
--ZZamboni
| [reply]
[d/l] |