Subj : Re: any function to handle this kind of counting?
To   : comp.programming
From : Rob Thorpe
Date : Wed Aug 03 2005 02:28 am

Peter Ammon wrote:
> Just wait 'til the Perl guy gets here, though!

I can take on that mantle:

for(0..($#a<$#b?$#a:$#b)){++$ans{$b[$_]}{$a[$_]};}

This is probably close to the shortest solution that can be done in any
language - but it's horrible.  Like all the other "terse" solutions it
misses the point.  In a real program you'll have to document exactly
why you're doing this operation and how.  Doing so will take a
significant number of lines.

For example documenting the perl:

# Now we have two arrays @a and @b we need to count each
# char-interger pair across these two arrays.  For example in:
# @a = ('a', 'b', 'a', 'e', 'a');
# @b = (  1,   2,   1,   4,   5); there are the pairs
# (1 ('a' 2))   (2 ('b' 1))   (4 ('e' 1))   (5 ('a' 1))
for ($i) (0..(($#a < $#b) ? $#a : $#b))  { # Find smallest array
    # Make %ans a hash-of-hashes containing the answer
    ++$ans{$b[$i]}{$a[$i]};
}

The ML and python versions people have presented also need much more
documentation.
The difference in terseness then becomes much smaller.  The difference
in # of lines likely to contain error is not so different though.

.