Match whole word in string including special characters

Question

I am aware of multiple existing answers that suggest:

def contains(string, word):
    return bool(re.search(rf"\b{word}\b", string))

But this pattern gives special treatment to alphanumeric character. For examples, contains("hello world!", "world!") returns False while contains("hello world!", "world") returns True.

I need a more 'naive' search pattern, one that matches a substring as long as it starts and ends with either the superstring's boundary or a space. (Desired behavior: opposite of the examples above.)

@anubhava that won't escape special characters and `\b` is the same as `(?<!\w)` — mousetail, May 16 '23 at 06:32
`\b is the same as (?<!\w)` not true but yes `re.search(rf"(?<!\w){word}(?!\w)", re.escape(string))` would be more safe — anubhava, May 16 '23 at 06:36
@Felix: Do you want to return `False` for `contains("hello world!", "world")`? — anubhava, May 16 '23 at 06:40

score 2 · Accepted Answer · edited May 17 '23 at 00:04

2

You need to avoid using \b (word boundary) and assert that previous and next positions don't have a non-whitespace character. Also it is safer to use re.escape as your search word may contain special regex meta characters.

You may use this python code:

def contains(string, word):
    return bool(re.search(rf"(?<!\S){word}(?!\S)", re.escape(string)))

print (contains("hello world!", "world"))
print (contains("hello world!", "world!"))

Output:

False
True

Online Code Demo

edited May 17 '23 at 00:04

Unmitigated

76,500
11
62
80

answered May 16 '23 at 06:45

anubhava

761,203
64
569
643

Is it necessary to `re.escape(word)` as well? (ref. @mousetail's comment) – Felix Fourcolor May 16 '23 at 06:49
1

Not necessary for this example but in case you have search word such as `wor(d` (containing regex meta characters) then yes it will be required. – anubhava May 16 '23 at 06:50

Match whole word in string including special characters

1 Answers1