Perl PDL，计算两个浮点值的出现次数

Question

我想同时计算两个浮点值的出现次数。

my @Key_Str = swcols "%0.2f_%0.2f", $a_datas, $b_datas;
my %counts;
foreach my $key_str (@Key_Str) {
    $counts{$key_str}++;
}
my (@p_a, @p_b @cnts);
foreach (sort keys %counts){
    my ($z, $h) = split /_/, $_;
    push @p_a, $z; push @p_b, $h; push @cnts, $counts{$_};
}
my $a= pdl(@p_a); my $b = pdl(@p_b); my $count = pdl(@cnts);

纯 PDL 是否可能实现同样的效果？

使用哈希和数组需要时间。

更详细，

#!/usr/bin/perl

use PDL;
use Data::Dumper;
use feature 'say';

my $a_data = pdl(1.5,  2.1,  2.1,  1.5,  -5.7);
my $b_data = pdl(12.2, 22.3, 22.3, 12.2, 15.3);

my @Key_Str = swcols "%0.2f_%0.2f", $a_data, $b_data;

my %counts;
foreach my $key_str (@Key_Str) {
    $counts{$key_str}++;
}
my (@p_a, @p_b, @cnts);
foreach (sort keys %counts){
    my ($z, $h) = split /_/, $_;
    push @p_a, $z; push @p_b, $h; push @cnts, $counts{$_};
}
my $a= pdl(@p_a); my $b = pdl(@p_b); my $count = pdl(@cnts);
say $a;
say $b;
say $count;

结果是

[-5.7 1.5 2.1]
[15.3 12.2 22.3]
[1 2 2]

Answer 1

我不确定如何使用简单的 PDL 函数获取唯一元素的计数，但您可以在循环中执行此操作。例如：

use v5.38;
use PDL;

# Sample values for $a_datas and $b_datas
my @a_datas = (1.2345, 2.3456, 1.2399, 3.4567, 2.3411, 1.2300, 3.4500, 4.5600, 2.3456, 1.2345);
my @b_datas = (3.4567, 4.5678, 3.4512, 5.6789, 4.5634, 3.4000, 5.6700, 6.7800, 4.5678, 3.4567);

my $a = pdl(@a_datas);
my $b = pdl(@b_datas);

# Round to two decimal places
$a = ($a * 100)->rint / 100;
$b = ($b * 100)->rint / 100;

# Combine into a 2D piddle
my $A = $a->cat($b);
$A = $A->xchg(0,1)->qsortvec;
my $B = $A->uniqvec;
say $B;

my $C = zeroes($B->dim(1));  # Initialize $C with zeros, with the same length as the number of rows in $B
my $j = 0;
for my $i (0 .. $B->dim(1) - 1) {
    my $Bval = $B->slice(":,($i)");
    my $count = 0;
    while (1) {
        my $Aval = $A->slice(":,($j)");
        if ($j < $A->dim(1) && all($Aval == $Bval)) {
            $count++;
            $j++;
        }
        else {
            last;
        }
    }
    $C->set($i, $count);
}
say $C;

输出：

[
 [1.23  3.4]
 [1.23 3.46]
 [1.24 3.45]
 [2.34 4.56]
 [2.35 4.57]
 [3.45 5.67]
 [3.46 5.68]
 [4.56 6.78]
]

[1 2 1 1 2 1 1 1]