Welcome to the Monastery!

I think you're over complicating it with all the erroneous double-quotes. If it's just plain ascii, try this:

If the d" is after a whitespace (ie. the beginning of a new word), you can do it like this to make the regex a bit more resilient:

The \b is a "word boundary", and using it in the position I did means 'if a literal lowercase d is followed by a double-quote and is at the beginning of a word (ie. following any whitespace)', replace the d" with an empty string.

Something that I just recently learned that was available and now use periodically besides YAPE::Regex::Explain is use re 'debug';. It might take a bit to get familiar with the output, but exceptionally valuable to learn if you want to understand why your regexes aren't (or are) matching when you think the opposite should be happening.

-stevieb

'd"alice' hardly makes any sense in French (which happens to be my mother tongue), at least not without a much more elaborate context. You could expect "d'alice" to indicate something belonging or pertaining to a given person named Alice, but I can't see a sense to 'd"alice'.

G'day nmdahl,

Welcome to the Monastery.

Here's a fairly straightforward way of doing that with a variety of test cases. Note this won't remove the 'd"' from the end of your first posted sentence (1st test case).

#!/usr/bin/env perl

use strict;
use warnings;

my $re = qr{\bd"(\w+)};

while (<DATA>) {
    s/$re/$1/g;
    print;
}

__DATA__
... and am having trouble with strings containing d".
... if the input string is d"alice I need ...
No d + " + alice here.
Here's many: d"alice and d"alice and d"alice and ...
d"alice at start of line
and at end of line: d"alice
"d"alice"
'd"alice'
[download]

Output:

... and am having trouble with strings containing d".
... if the input string is alice I need ...
No d + " + alice here.
Here's many: alice and alice and alice and ...
alice at start of line
and at end of line: alice
"alice"
'alice'
[download]

If that doesn't do what you want, please post representative input and expected output. The guidelines in "How do I post a question effectively?" explain the best way to do that.

Encoding. What is the text file encoding?

Thank you for the help everyone! stevieb your solution was exactly what I was looking for. Just for everyone's reference, an example input would be appelle la maison d"Alice. The output in this case would be appelle maison Alice. I have had no problems removing la and similar words, but was having some difficulty with d" since it was part of a word I wanted to keep. But it is working now, thank you everyone.

Once again, you would never see a sentence such as 'appelle la maison d"Alice' in French. Or only as an improbable error, because this is just plainly incorrect. In French, double quotes or guillemets are just the same thing as what they are in English and most European languages, i.e. a quoting punctuation character, not a part of the language itself.

What you could see in a French sentence is "appelle la maison d'Alice", with a single quote (or apostrophe) between the "d" letter and the word "Alice".

So your use case still does not make any sense to me.