Markus Jais <mjais / web.de> writes:
> but this does not work. I thought that if this would work
> it could be faster because the block is only called 
> for the keys which macht the regexp.

I don't think there will be any improvement. Assuming that the syntax
hash[/^b/].each works, then each element still has to be visited for
matching. So, this moves the matching/checking from within the block
to within the hash.each method. The only work saved is the call to
method 'yield' from within #each. Total work is still largely the
same: the computer has to check each element of the hash.

If you have a large number of elements such that the overhead of
performing yield to the block becomes significant, then most probably
you should have refactored/restructure the way you store things.

YS.