Question

Better explained with examples:

1. HHH
2. HHHH
3. HHHBBHHH
4. HHHBH
5. BB
6. HHBH

I need to come up with a regexp that matches only 3 H’s or a multiple of 3 H’s (so 6, 9, 12, … H’s are ok as well) and 5 H’s are not ok. And if possible I don’t want to use Perl regexps.

So for the input above the regexp would match (1), (3) and (6) only.
I’m just starting with regular expressions here so I don’t exactly know how I’m supposed to approach this.

edit
Just to clear something up:, an H can only be in one group of 3 H’s. The group of 3 H’s might be HHH or HHBH.
That’s why in example 2 above it is not a match because the last H is not in a group of 3 H’s. And you can’t take the last 3 H’s in a group because the middle 2 H’s have already been inside a group before.

Answer 1

You can use the following regular expression:

^([^H]*H[^H]*H[^H]*H[^H]*)+$

It matches any string which contains in total 3 H or any multiple of 3. In between there might be any other character.

Explanation:

^              begin of string
(              start of group
  [^H]*H       any string of characters (or none) not including 'H' plus a single 'H'
  [^H]*H       any string of characters (or none) not including 'H' plus a single 'H'
  [^H]*H       any string of characters (or none) not including 'H' plus a single 'H'
  [^H]*        any string of characters (or none) which is not 'H'
)+             containing the group once or twice or ...
$              end of string

By repeating the subpattern [^H]*H three times we make sure that there are indeed 3 H included, [^H]* allows any separating characters.

Note: use either egrep or run grep with additional argument -E.