[Nh-pm] The Perly Way

[Paul L Lussier]

In a message dated: Wed, 14 May 2003 14:48:55 EDT
Ray Cote said:

>>  ... so I would have gone after the efficiency of the the third

>To expand my Perl knowledge a little bit more, do you feel the third 
>example is more efficient because it does not have the overhead of 
>building a second array?

In short yes.   The third is nothing more than a refinement of the 
second.  Technically, the second probably uses slightly less memory, 
since it just spits out each element as it appears.  However, not 
wanting a stray \t at the end of every line (which could introduce 
line-wrapping issues) I'd rather incur the overhead of storing 
everything in $string, then printing each complete row out all at 
once, rather than in pieces.

>However, what is the efficiency of always appending to the $string on 
>each pass vs doing a join (though, for all I know, join may simply 
>iterate and append as well).

I have no idea what the difference between .= and a join are as far 
as efficiency.  The .= though, allows you to operate on each element 
before printing it, whereas a join does not.  Really, I think it's a 
wash in this case.  Any of the 3 methods is slightly more efficient 
than the pre-processing you started out with.  I don't by how much, 
since it completely depends upon your dataset.  But having to process 
each row twice, basically adds an exponential amount of time your 
processing, right?

>>  (which incidently, has a
>>minor bug in it, can you spot it :)
>You didn't declare my $string?

Well, declaration is irrellevent in perl in most cases.  More 
specifically, I do not clear the contents of $string after each pass 
of the loop.  So, since I use the .= operator, I end up appending 
each row to the next for the duration of the loop.

This problem can be solved in 2 ways.  Either, create $string as a 
my()'ed variable local to the foreach, so that it remains in scope 
for the map, but gets re-defined for each new row, or just clear the 
variable at the beginning of each pass of the foreach loop.  In 
either case, the statement would just before the map() call.

Clearing the variable probably is the most efficient, since you only 
allocate that memory once, where as my()'ing the variable probably 
allocates a new chunk of memory through each pass.  The difference in 
performance is probably negligible, but, it's good to be aware of the 
implications of different methods :)
-- 

Seeya,
Paul
--
Key fingerprint = 1660 FECC 5D21 D286 F853  E808 BB07 9239 53F1 28EE

	It may look like I'm just sitting here doing nothing,
   but I'm really actively waiting for all my problems to go away.

	 If you're not having fun, you're not doing it right!