Regex term for batch identifying word count in comments

[slantybard]

I am trying to clean up my comments and want to remove word counts from some of them. For example, I have some books where the end of the comments have:

"x,xxx Words" or "xx,xxx Words" or "xxx,xxx Words"

I know I can use (\d+) to select all numbers....but....this also selects any other random numbers, dates, or chapter numbers that might be present as well.

Is there a specific regex that would only select the phrase "x,xxx Words" or "xx,xxx Words" or "xxx,xxx Words" where the x=any digit?

A specific example of what I'm looking for would be seen in this book comment:

This is an extremely AU crossover fic that asks the question what might have happened if Petunia Dursley hadn't found a young Harry Potter sleeping on her doorstep on the morning of the 2nd of November 1981. After all, Dumbledore was a bit careless with him 109,446 Words

Here, I want to use a regex to delete the "109,446 Words" but leave the "2nd" and "1981" alone.

I don't mind splitting up the regex into several searches depending on number of words. Any suggestions or links to where I can find an answer would be appreciated.

[ownedbycats]

While there's probably a better way to handle it, this will match xxx,xxx words (or Words):

https://regex101.com/r/YhHpr3/1

For other digits, just edit the two {3} bits to reflect the number of digits.

EDIT: Here's an improved version that can handle other numbers of digits:

https://regex101.com/r/YhHpr3/2

[davidfor]

Slightly better would be:

(\d{0,3}[,]?\d{1,3} [Ww]ords)

That should get any number up to 6 digits with or without the comma. But, it would miss the first digit in 1000000.

If I hadn't see your solution, I think I would have suggested:

([\d,]+ [Ww]ords)

Which would pickup ", words". So, slightly safer is:

([\d,]{2,} [Ww]ords\b)

[slantybard]

Thank you both very much for the suggestions. Regex is wonderfully wacky! I ended up using the original suggestion prior to seeing the improved suggestions. In the end, I used:

(\d{1}[,]\d{3}[,]\d{3} (W|w)ords)

To remove all the comments in books with > 1,000,000 words

(\d{3}[,]\d{3} (W|w)ords)

To remove all the comments in books between 100,000 and 1,000,000 words

(\d{2}[,]\d{3} (W|w)ords)

To remove all the comments in books between 10,000 and 100,000 words

(\d{1}[,]\d{3} (W|w)ords)

To remove all the comments in books between 1,000 and 10,000 words

(\d{3} (W|w)ords)

To finally remove all the comments in books < 1,000 words

[davidfor]

A single one to do this is:

(\d{0,1}[,]?\d{0,3}[,]?\d{1,3} [Ww]ords)

It isn't perfect because it would pickup something like "1,,1 Words". But, I suspect that won't appear.

And semantically, "[Ww]" is a bit better than "(W|w)". The former is "One of this list of things" and the later is "W or w and put them into a group for the replacement". It is a little pedantic and doesn't make a difference in your situation. But, if you need to extend this because you found another case, or, needed to replace the text, the latter might not work as well.

[ownedbycats]

Yes, I'd forgotten that square brackets worked better for that.