Find all matches with an extra qualifying criteria

Question

Given sentences such as;

Boy has a dog and a cat.
Boy microwaves a gerbil.
Sally owns a cat.

For each sentence I want a list of animals (de

Answer 1

You may use a positive lookbehind:

(?<=^Boy.*?)(?:dog|cat|gerbil)

Or, a variation with word boundaries to match the animals as whole words:

(?<=^Boy\b.*?)\b(?:dog|cat|gerbil)\b

See the regex demo

The (?<=^Boy.*?) positive lookbehind will require the Boy at the start of the string for the consuming pattern to match.

If your input contains LF (newline) chars, pass the RegexOptions.Singleline option for . to match newlines, too.

C# usage:

var results = Regex.Matches(s, @"(?<=^Boy\b.*?)\b(?:dog|cat|gerbil)\b")
        .Cast<Match>()
        .Select(m => m.Value)
        .ToList();

C# demo:

var strs = new List<string>() { "Boy has a dog and a cat.", 
        "Boy something a gerbil.",
        "Sally owns a cat." };
foreach (var s in strs)
{
    var results = Regex.Matches(s, @"(?<=^Boy\b.*?)\b(?:dog|cat|gerbil)\b")
            .Cast<Match>()
            .Select(m => m.Value)
            .ToList();
     if (results.Count > 0) {
        Console.WriteLine("{0}:\n[{1}]\n------", s, string.Join(", ", results));
     }
     else
     {
        Console.WriteLine("{0}:\nNO MATCH!\n------", s);
     }
}

Output:

Boy has a dog and a cat.:
[dog, cat]
------
Boy something a gerbil.:
[gerbil]
------
Sally owns a cat.:
NO MATCH!
------

There is an alternative: match any string starting with Boy and then after each successful match only:

(?:\G(?!\A)|^Boy\b).*?\b(dog|cat|gerbil)\b

See this regex demo (or a regex101 link here)

You would just need to grab Group 1 contents:

var results = Regex.Matches(s, @"(?:\G(?!\A)|^Boy\b).*?\b(dog|cat|gerbil)\b")
            .Cast<Match>()
            .Select(m => m.Groups[1].Value)
            .ToList();

See this C# demo.

Here,

• (?:\G(?!\A)|^Boy\b) - either the end of the precvious match (\G(?!\A)) or the start of the string followed with the whole word Boy
• .*? - any 0+ chars other than a newline (if no RegexOptions.Singleline is passed to the Regex constructor) as few as possible
• \b(dog|cat|gerbil)\b - a whole word dog, cat or gerbil

Bascially, these regexps are similar, although \G based regex might turn out a bit faster.