Beefy Boxes and Bandwidth Generously Provided by pair Networks
Perl-Sensitive Sunglasses
 
PerlMonks  

Re: Fast seeking in a large array of hashes to generate a report.

by themage (Friar)
on Jun 23, 2005 at 11:34 UTC ( #469364=note: print w/ replies, xml ) Need Help??


in reply to Fast seeking in a large array of hashes to generate a report.

Hi ppl, Hi jbrugger, I created a small script that generates an array with 1.000.000 records as listed in your example as follow:

use strict; use Time::HiRes qw(time); my @data=(); my @names=qw(albert arthur bernard bill charlie david jack jonh joseph + mark michael peter steven); my @surnames=qw(bell brown canfield cornwell devlin doyle golden hoffm +an maclean powell rowling tovard twain warlick); my @places=qw(amsterdan athens belfast berlin bern brussels copenhagen + helsinki lisbon london luxenbourg madrid oslo paris rome stockholm v +aduz vienna); for (my $i=0;$i<=1000000;$i++) { my $name=$names[int(rand()*$#names)]; my $surname=$surnames[int(rand()*$#surnames)]; my $place=$places[int(rand()*$#places)]; my $age=int(rand()*50)+25; my $rec={name=>$name,surname=>$surname,place=>$place,age=>$age +}; push @data, $rec; }
Copied @data to @Persons and @Person2, and used the @Filters arrays as follow:
my @Fields=( { key => 'age', content=>35 }, { key => 'place', content=>'london'}, ); my @Persons=@data; my @Person2=@data; print "Persons: $#Persons\nPerson2: $#Person2\n\n"; my $field; my $person; my $i=0;
Timed your implementation with:
my $ta=time(); foreach $field (@Fields) { # <- age and surname in this case scalar(@Persons)==0 && last; for ($i=0;$i<scalar(@Persons);$i++) { ($Persons[$i]->{$field->{key}} eq $field->{content}) && next; splice(@Persons,$i,1); $i--; } } my $tb=time();
And then made a small change in your filtering algorith, so the main interaction is against the data, not against the filters as follow:
my $use=1; my @Rset=(); foreach $person (@Person2) { $use=1; foreach $field (@Fields) { if ($person->{$field->{key}} ne $field->{content}) { $use=0; last; } } if ($use) { push @Rset, $person; } } my $tc=time;
And then printed some stats:
my $iab=$tb-$ta; my $ibc=$tc-$tb; print "TIMES:\nab: $iab\nbc: $ibc\n"; print "COUNTS:\nab: ", $#Persons,"\nbc: ",$#Rset,"\n";
I was expecting some diference (not know better or worst, was just a test), and got some. Enexpected, I admit. In my P4 3.2Ghz, 1GB memory this are some results:
mpneves@voyager perl$ ./test.pl Persons: 1000000 Person2: 1000000 TIMES: ab: 17.5640239715576 bc: 1.96199607849121 COUNTS: ab: 1153 bc: 1153 mpneves@voyager perl$ ./test.pl Persons: 1000000 Person2: 1000000 TIMES: ab: 18.5665609836578 bc: 1.95595502853394 COUNTS: ab: 1185 bc: 1185 mpneves@voyager perl$ ./test.pl Persons: 1000000 Person2: 1000000 TIMES: ab: 24.4221110343933 bc: 3.58428502082825 COUNTS: ab: 1213 bc: 1213
Which means that my version is performing better at least 1 to 3 than yours. Obviously, this is a test in my notebok, with random data. with your computers, with real data and filters it can performe diferently. All code I used is here. If you join all code segments as they are presented you will have my complete code.


Comment on Re: Fast seeking in a large array of hashes to generate a report.
Select or Download Code

Log In?
Username:
Password:

What's my password?
Create A New User
Node Status?
node history
Node Type: note [id://469364]
help
Chatterbox?
and the web crawler heard nothing...

How do I use this? | Other CB clients
Other Users?
Others wandering the Monastery: (12)
As of 2015-07-06 21:11 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 (83 votes), past polls