Count the number of words ends with specific character using Bash

Question

Heys guys I wanted to count the number of words that end with 's' from this URL,

https://matt.might.net/articles/what-cs-majors-should-know/

Here's what I did

curl https://matt.might.net/articles/what-cs-majors-should-know/ | \
grep s$ 

I couldn't find a way to filter out the words that end with 's' and count them

Answer 1

elinks -no-numbering -no-references -dump https://matt.might.net/articles/what-cs-majors-should-know/ |
  grep -Po '\w+s\b' |
  wc -l

(with GNU grep or compatible)

Gives me: 595

elinks retrieves the HTML and converts it to text in its own way. Some alternatives are w3m -dump or lynx -nolist.

Then we look for the sequences of one or more word characters in there (word characters being alnums or underscores) followed by a s provided it's not followed by a word character, and count then with wc -l.

In essence, the definition of a word there is a sequence of 2 or more alnum or underscore characters and we look for the ones ending in s (add a -i option to grep to also look for those ending in S). That means the s in it's for instance is not counted. If you want wanted to consider single-letter words, you could replace + with * above. Or just do grep -Po 's\b' but then you wouldn't be able to see the list of words when omitting the | wc -l.

Another approach would be to extract the words first and find the ones ending in s:

grep -Eo '\w+' | grep -c 's$'

In any case, that's quite a crude definition of word that only works properly for simple English text. When applied to text in other languages, you could run into problems. For instance, it would find a match inside the abrogowałybyście Polish word (picked at random in a word list) if that ś was encoded in its decomposed sU+0301 form (where U+0301 is the combining acute accent) as that s would be seen as not being followed by a word character.

You may also want to consider what to do about the S in USA or U.S.A..