A question on regexprep

This is a question stem from a Cody Problem when I looked through those efficient solutions. Let's say I have the following code:

s = 'John got an A in math.';
regexprep(s,'\<(\w)(\w*)(\w)\>','$1${$2(end:-1:1)}$3');

If I try to replace each match with only one of its tokens:

regexprep(s,'\<(\w)(\w*)(\w)\>','$1'); % or '$2','$3'

I will find that 'A' can be recognized as any one of them, and yet when it is replaced by '$1${$2(end:-1:1)}$3', only one copy of 'A' is returned. Why?

Thank you in advance.

Answer 1

[TADA]

A is not a match for your regexp,

match = regexp(s,'\<(\w)(\w*)(\w)\>', 'match')

match =

  1×5 cell array

    {'John'}    {'got'}    {'an'}    {'in'}    {'math'}
    

so it's not replaced by anything,

1. 'John' was replaced by 'ho'
2. 'got' was replaced by 'o'
3. 'an' was replaced by ''
4. A is not a match, not replaced
5. 'in' was replaced by ''
6. 'math' was replaced by 'ta'

    s2 = regexprep(s,'\<(\w)(\w*)(\w)\>','${$2(end:-1:1)}')

s2 =

    'ho o  A  ta.'

[Edward Huang]

Oops... I didn't realize I could have checked with regexp... Thanks!