PHP优化20万+PDF文件按修改日期分组统计性能问题求助

我需要从目录/pdf_files/dir中获取PDF文件列表，生成按修改日期分组统计的HTML报告。但该目录包含20万+文件，当前三种实现方案运行速度极慢：

尝试1：使用glob与filemtime（耗时42秒） #

<?php
$array_ids = glob('/pdf_files/dir/*.pdf',GLOB_BRACE);
foreach ( $array_ids as $x)
{
    $array2[] = array($x, filemtime($x));
}

$count_array = array();
foreach ($array2 as $x)
{
    $mon = date ("m", filemtime($x[0]));
    $mon2 = date ("M", filemtime($x[0]));
    $year = date ("Y", filemtime($x[0]));
    ${'count_' . $year . '_' . $mon. '_' . $mon2} = ${'count_' . $year . '_' . $mon. '_' . $mon2} + 1 ;

    if (! in_array("count_${year}_${mon}_${mon2}" , $count_array ) )
    {
        $count_array[] = "count_${year}_${mon}_${mon2}" ;
    }

}
sort($count_array);

echo "<table border=1>";

foreach ( $count_array as $x)
{
    $d_x = substr($x,-3) . ' ' . substr($x,6,4) ;
    echo "<tr><td style='text-align:center; vertical-align:middle; padding: 15px;'>$d_x</td><td style='text-align:center; vertical-align:middle; padding: 15px;'>${$x}</td></tr>";
    $total = $total + $$x ;
}
echo "<tr><td style='text-align:center; vertical-align:middle; padding: 15px;'>Total</td><td style='text-align:center; vertical-align:middle; padding: 15px;'>$total</td></tr></table>";
?>

尝试2：使用stat（耗时37秒） #

<?php
$count_array = array();

for ( $i = 1 ; $i < 10 ; $i++ )
{
    $output = shell_exec("cd /pdf_files/dir ; stat -c '%n,%Y' $i*.pdf");
    if ($output != "" )
    {
        $array2[] = explode("\n", $output);
    }
}
$array2 = call_user_func_array('array_merge', $array2);
$array2 = array_filter($array2);

foreach ($array2 as $x)
{
    $filename = explode(",", $x)[0];
    $filetime = explode(",", $x)[1];
    $mon = date ("m", $filetime);
    $mon2 = date ("M", $filetime);
    $year = date ("Y", $filetime);
    ${'count_' . $year . '_' . $mon. '_' . $mon2} = ${'count_' . $year . '_' . $mon. '_' . $mon2} + 1 ;

    if (! in_array("count_${year}_${mon}_${mon2}" , $count_array ) )
    {
        $count_array[] = "count_${year}_${mon}_${mon2}" ;
    }

}
sort($count_array);

echo "<table border=1>";

foreach ( $count_array as $x)
{
    $d_x = substr($x,-3) . ' ' . substr($x,6,4) ;
    echo "<tr><td style='text-align:center; vertical-align:middle; padding: 15px;'>$d_x</td><td style='text-align:center; vertical-align:middle; padding: 15px;'>${$x}</td></tr>";
    $total = $total + $$x ;
}
echo "<tr><td style='text-align:center; vertical-align:middle; padding: 15px;'>Total</td><td style='text-align:center; vertical-align:middle; padding: 15px;'>$total</td></tr></table>";
?>

尝试3：使用ls（耗时33秒） #

<?php
$count_array = array();

for ( $i = 1 ; $i < 10 ; $i++ )
{
    $output = shell_exec("cd /pdf_files/dir ; ls -l --time-style=\"+%s\" $i*.pdf | awk '{print $7 \",\" $6}'");
    $array2[] = explode("\n", $output);
}
$array2 = call_user_func_array('array_merge', $array2);
$array2 = array_filter($array2);

foreach ($array2 as $x)
{
    $filename = explode(",", $x)[0];
    $filetime = explode(",", $x)[1];
    $mon = date ("m", $filetime);
    $mon2 = date ("M", $filetime);
    $year = date ("Y", $filetime);
    ${'count_' . $year . '_' . $mon. '_' . $mon2} = ${'count_' . $year . '_' . $mon. '_' . $mon2} + 1 ;

    if (! in_array("count_${year}_${mon}_${mon2}" , $count_array ) )
    {
        $count_array[] = "count_${year}_${mon}_${mon2}" ;
    }

}
sort($count_array);

echo "<table border=1>";

foreach ( $count_array as $x)
{
    $d_x = substr($x,-3) . ' ' . substr($x,6,4) ;
    echo "<tr><td style='text-align:center; vertical-align:middle; padding: 15px;'>$d_x</td><td style='text-align:center; vertical-align:middle; padding: 15px;'>${$x}</td></tr>";
    $total = $total + $$x ;
}
echo "<tr><td style='text-align:center; vertical-align:middle; padding: 15px;'>Total</td><td style='text-align:center; vertical-align:middle; padding: 15px;'>$total</td></tr></table>";
?>

预期输出示例 #

May 2024    8852
Jun 2024    14108
Jul 2024    15965
Aug 2024    18139
Sep 2024    13920
Oct 2024    14625
Nov 2024    17029
Dec 2024    39909
Total   142547

我曾尝试使用ls -ltr但无法完成循环与解析，现寻求高效优化方案。

内容的提问来源于stack exchange，提问作者huz_akh