Beefy Boxes and Bandwidth Generously Provided by pair Networks
Clear questions and runnable code
get the best and fastest answer
 
PerlMonks  

Re: Concerning hash operations (appending, concatenating)

by inq123 (Sexton)
on Mar 24, 2005 at 18:20 UTC ( #442145=note: print w/ replies, xml ) Need Help??


in reply to Concerning hash operations (appending, concatenating)

Note that %combined = (%hash1, %hash2); suffers from performance penalty. If performance's a concern, it's always better to do map { $hash1{$_} = $hash2{$_} } keys %hash2; (unless you want to keep %hash1 and %hash2 intact, which doesn't seem to be the case in your question)

use Benchmark; my $size = 100000; my %hash1 = map { $_ => 1 } (0..$size); my %hash2 = map { $_ => 1 } ($size..(2*$size)); my $t0 = new Benchmark; my %combined = (%hash1, %hash2); my $t1 = new Benchmark; print "Combining took:",timestr(timediff($t1, $t0)),"\n"; $t0 = new Benchmark; map { $hash1{$_} = $hash2{$_} } keys %hash2; my $t1 = new Benchmark; print "Map took:",timestr(timediff($t1, $t0)),"\n";
When size = 100000 and 1000000 respectively, the results:
inq123@perlmonks$ perl test.pl
Combining took: 0 wallclock secs ( 0.28 usr +  0.02 sys =  0.30 CPU)
Map took: 0 wallclock secs ( 0.18 usr +  0.00 sys =  0.18 CPU)
inq123@perlmonks$ perl test.pl
Combining took:42 wallclock secs (41.53 usr +  0.21 sys = 41.74 CPU)
Map took: 2 wallclock secs ( 1.89 usr +  0.05 sys =  1.94 CPU)
So performance penalty is manifested when the hashs contain tens of thousands of elements, which is not too rare.


Comment on Re: Concerning hash operations (appending, concatenating)
Download Code
Re^2: Concerning hash operations (appending, concatenating)
by tilly (Archbishop) on Mar 26, 2005 at 02:10 UTC
    General advice. Just because there are many ways to do it in Perl is not a reason for picking a less readable one. If you are not returning data from map, then you should write it like this:
    $hash1{$_} = $hash2{$_} for keys %hash2;
    That is easier to read, and much more clearly signals intent. It is also at least as fast as the map version. (It used to be a lot faster, but in Perl 5.8 there is an optimization that causes map to shortcircuit to become a for if it is in null context.)

    Furthermore performance is far less likely to matter than most people think, and when it does having micro-optimized as you went is generally a bad strategy for getting it. (You want to keep code clean and then look for a better algorithm, or move a small section into C.) Therefore I would generally use the following strategy because it is even clearer, even though it is marginally slower on my machine (about 10% so):

    @hash1{keys %hash2} = values %hash2;
    And, of course, in the rare case that performance really mattered and I really wanted to work in Perl, it is fastest to avoid having to do 2 sets of hash lookups on %hash2:
    $combined{$k} = $v while my ($k, $v) = each %hash2;

Log In?
Username:
Password:

What's my password?
Create A New User
Node Status?
node history
Node Type: note [id://442145]
help
Chatterbox?
and the web crawler heard nothing...

How do I use this? | Other CB clients
Other Users?
Others surveying the Monastery: (8)
As of 2015-07-03 11:03 GMT
Sections?
Information?
Find Nodes?
Leftovers?
    Voting Booth?

    The top three priorities of my open tasks are (in descending order of likelihood to be worked on) ...









    Results (51 votes), past polls