[php] How to get directory size in PHP

function foldersize($path) {
  $total_size = 0;
  $files = scandir($path);

  foreach($files as $t) {
    if (is_dir(rtrim($path, '/') . '/' . $t)) {
      if ($t<>"." && $t<>"..") {
          $size = foldersize(rtrim($path, '/') . '/' . $t);

          $total_size += $size;
      }
    } else {
      $size = filesize(rtrim($path, '/') . '/' . $t);
      $total_size += $size;
    }
  }
  return $total_size;
}

function format_size($size) {
  $mod = 1024;
  $units = explode(' ','B KB MB GB TB PB');
  for ($i = 0; $size > $mod; $i++) {
    $size /= $mod;
  }

  return round($size, 2) . ' ' . $units[$i];
}

$SIZE_LIMIT = 5368709120; // 5 GB

$sql="select * from users order by id";
$result=mysql_query($sql);

while($row=mysql_fetch_array($result)) {
  $disk_used = foldersize("C:/xampp/htdocs/freehosting/".$row['name']);

  $disk_remaining = $SIZE_LIMIT - $disk_used;
  print 'Name: ' . $row['name'] . '<br>';

  print 'diskspace used: ' . format_size($disk_used) . '<br>';
  print 'diskspace left: ' . format_size($disk_remaining) . '<br><hr>';
}

php disk_total_space

Any idea why the processor usage shoot up too high or 100% till the script execution is finish ? Can anything be done to optimize it? or is there any other alternative way to check folder and folders inside it size?

This question is related to php filesystems

The answer is


Even though there are already many many answers to this post, I feel I have to add another option for unix hosts that only returns the sum of all file sizes in the directory (recursively).

If you look at Jonathan's answer he uses the du command. This command will return the total directory size but the pure PHP solutions posted by others here will return the sum of all file sizes. Big difference!

What to look out for

When running du on a newly created directory, it may return 4K instead of 0. This may even get more confusing after having deleted files from the directory in question, having du reporting a total directory size that does not correspond to the sum of the sizes of the files within it. Why? The command du returns a report based on some file settings, as Hermann Ingjaldsson commented on this post.

The solution

To form a solution that behaves like some of the PHP-only scripts posted here, you can use ls command and pipe it to awk like this:

ls -ltrR /path/to/dir |awk '{print \$5}'|awk 'BEGIN{sum=0} {sum=sum+\$1} END {print sum}'

As a PHP function you could use something like this:

function getDirectorySize( $path )
{
    if( !is_dir( $path ) ) {
        return 0;
    }

    $path   = strval( $path );
    $io     = popen( "ls -ltrR {$path} |awk '{print \$5}'|awk 'BEGIN{sum=0} {sum=sum+\$1} END {print sum}'", 'r' );
    $size   = intval( fgets( $io, 80 ) );
    pclose( $io );

    return $size;
}

function get_dir_size($directory){
    $size = 0;
    $files = glob($directory.'/*');
    foreach($files as $path){
        is_file($path) && $size += filesize($path);
        is_dir($path)  && $size += get_dir_size($path);
    }
    return $size;
} 

if you are hosted on Linux:

passthru('du -h -s ' . $DIRECTORY_PATH)

It's better than foreach


PHP get directory size (with FTP access)

After hard work, this code works great!!!! and I want to share with the community (by MundialSYS)

function dirFTPSize($ftpStream, $dir) {
    $size = 0;
    $files = ftp_nlist($ftpStream, $dir);

    foreach ($files as $remoteFile) {
        if(preg_match('/.*\/\.\.$/', $remoteFile) || preg_match('/.*\/\.$/', $remoteFile)){
            continue;
        }
        $sizeTemp = ftp_size($ftpStream, $remoteFile);
        if ($sizeTemp > 0) {
            $size += $sizeTemp;
        }elseif($sizeTemp == -1){//directorio
            $size += dirFTPSize($ftpStream, $remoteFile);
        }
    }

    return $size;
}

$hostname = '127.0.0.1'; // or 'ftp.domain.com'
$username = 'username';
$password = 'password';
$startdir = '/public_html'; // absolute path
$files = array();
$ftpStream = ftp_connect($hostname);
$login = ftp_login($ftpStream, $username, $password);
if (!$ftpStream) {
    echo 'Wrong server!';
    exit;
} else if (!$login) {
    echo 'Wrong username/password!';
    exit;
} else {
    $size = dirFTPSize($ftpStream, $startdir);
}

echo number_format(($size / 1024 / 1024), 2, '.', '') . ' MB';

ftp_close($ftpStream);

Good code! Fernando


It works perfectly fine .

     public static function folderSize($dir)
     {
        $size = 0;

        foreach (glob(rtrim($dir, '/') . '/*', GLOB_NOSORT) as $each) {
            $func_name = __FUNCTION__;
            $size += is_file($each) ? filesize($each) : static::$func_name($each);
        }

        return $size;
      }

Just another function using native php functions.

function dirSize($dir)
    {
        $dirSize = 0;
        if(!is_dir($dir)){return false;};
        $files = scandir($dir);if(!$files){return false;}
        $files = array_diff($files, array('.','..'));

        foreach ($files as $file) {
            if(is_dir("$dir/$file")){
                 $dirSize += dirSize("$dir/$file");
            }else{
                $dirSize += filesize("$dir/$file");
            }
        }
        return $dirSize;
    }

NOTE: this function returns the files sizes, NOT the size on disk


I found this approach to be shorter and more compatible. The Mac OS X version of "du" doesn't support the -b (or --bytes) option for some reason, so this sticks to the more-compatible -k option.

$file_directory = './directory/path';
$output = exec('du -sk ' . $file_directory);
$filesize = trim(str_replace($file_directory, '', $output)) * 1024;

Returns the $filesize in bytes.


PHP get directory size (with FTP access)

After hard work, this code works great!!!! and I want to share with the community (by MundialSYS)

function dirFTPSize($ftpStream, $dir) {
    $size = 0;
    $files = ftp_nlist($ftpStream, $dir);

    foreach ($files as $remoteFile) {
        if(preg_match('/.*\/\.\.$/', $remoteFile) || preg_match('/.*\/\.$/', $remoteFile)){
            continue;
        }
        $sizeTemp = ftp_size($ftpStream, $remoteFile);
        if ($sizeTemp > 0) {
            $size += $sizeTemp;
        }elseif($sizeTemp == -1){//directorio
            $size += dirFTPSize($ftpStream, $remoteFile);
        }
    }

    return $size;
}

$hostname = '127.0.0.1'; // or 'ftp.domain.com'
$username = 'username';
$password = 'password';
$startdir = '/public_html'; // absolute path
$files = array();
$ftpStream = ftp_connect($hostname);
$login = ftp_login($ftpStream, $username, $password);
if (!$ftpStream) {
    echo 'Wrong server!';
    exit;
} else if (!$login) {
    echo 'Wrong username/password!';
    exit;
} else {
    $size = dirFTPSize($ftpStream, $startdir);
}

echo number_format(($size / 1024 / 1024), 2, '.', '') . ' MB';

ftp_close($ftpStream);

Good code! Fernando


Evolved from Nate Haugs answer I created a short function for my project:

function uf_getDirSize($dir, $unit = 'm')
{
    $dir = trim($dir, '/');
    if (!is_dir($dir)) {
        trigger_error("{$dir} not a folder/dir/path.", E_USER_WARNING);
        return false;
    }
    if (!function_exists('exec')) {
        trigger_error('The function exec() is not available.', E_USER_WARNING);
        return false;
    }
    $output = exec('du -sb ' . $dir);
    $filesize = (int) trim(str_replace($dir, '', $output));
    switch ($unit) {
        case 'g': $filesize = number_format($filesize / 1073741824, 3); break;  // giga
        case 'm': $filesize = number_format($filesize / 1048576, 1);    break;  // mega
        case 'k': $filesize = number_format($filesize / 1024, 0);       break;  // kilo
        case 'b': $filesize = number_format($filesize, 0);              break;  // byte
    }
    return ($filesize + 0);
}

There are several things you could do to optimise the script - but maximum success would make it IO-bound rather than CPU-bound:

  1. Calculate rtrim($path, '/') outside the loop.
  2. make if ($t<>"." && $t<>"..") the outer test - it doesn't need to stat the path
  3. Calculate rtrim($path, '/') . '/' . $t once per loop - inside 2) and taking 1) into account.
  4. Calculate explode(' ','B KB MB GB TB PB'); once rather than each call?

function get_dir_size($directory){
    $size = 0;
    $files = glob($directory.'/*');
    foreach($files as $path){
        is_file($path) && $size += filesize($path);
        is_dir($path)  && $size += get_dir_size($path);
    }
    return $size;
} 

Object Oriented Approach :

/**
 * Returns a directory size
 *
 * @param string $directory
 *
 * @return int $size directory size in bytes
 *
 */
function dir_size($directory)
{
    $size = 0;
    foreach(new RecursiveIteratorIterator(new RecursiveDirectoryIterator($directory)) as $file)
    {
        $size += $file->getSize();
    }
    return $size;
}

Fast and Furious Approach :

function dir_size2($dir)
{
    $line = exec('du -sh ' . $dir);
    $line = trim(str_replace($dir, '', $line));
    return $line;
}

Even though there are already many many answers to this post, I feel I have to add another option for unix hosts that only returns the sum of all file sizes in the directory (recursively).

If you look at Jonathan's answer he uses the du command. This command will return the total directory size but the pure PHP solutions posted by others here will return the sum of all file sizes. Big difference!

What to look out for

When running du on a newly created directory, it may return 4K instead of 0. This may even get more confusing after having deleted files from the directory in question, having du reporting a total directory size that does not correspond to the sum of the sizes of the files within it. Why? The command du returns a report based on some file settings, as Hermann Ingjaldsson commented on this post.

The solution

To form a solution that behaves like some of the PHP-only scripts posted here, you can use ls command and pipe it to awk like this:

ls -ltrR /path/to/dir |awk '{print \$5}'|awk 'BEGIN{sum=0} {sum=sum+\$1} END {print sum}'

As a PHP function you could use something like this:

function getDirectorySize( $path )
{
    if( !is_dir( $path ) ) {
        return 0;
    }

    $path   = strval( $path );
    $io     = popen( "ls -ltrR {$path} |awk '{print \$5}'|awk 'BEGIN{sum=0} {sum=sum+\$1} END {print sum}'", 'r' );
    $size   = intval( fgets( $io, 80 ) );
    pclose( $io );

    return $size;
}

I found this approach to be shorter and more compatible. The Mac OS X version of "du" doesn't support the -b (or --bytes) option for some reason, so this sticks to the more-compatible -k option.

$file_directory = './directory/path';
$output = exec('du -sk ' . $file_directory);
$filesize = trim(str_replace($file_directory, '', $output)) * 1024;

Returns the $filesize in bytes.


Regarding Johnathan Sampson's Linux example, watch out when you are doing an intval on the outcome of the "du" function, if the size is >2GB, it will keep showing 2GB.

Replace:

$totalSize = intval(fgets($io, 80));

by:

strtok(fgets($io, 80), " ");

supposed your "du" function returns the size separated with space followed by the directory/file name.


There are several things you could do to optimise the script - but maximum success would make it IO-bound rather than CPU-bound:

  1. Calculate rtrim($path, '/') outside the loop.
  2. make if ($t<>"." && $t<>"..") the outer test - it doesn't need to stat the path
  3. Calculate rtrim($path, '/') . '/' . $t once per loop - inside 2) and taking 1) into account.
  4. Calculate explode(' ','B KB MB GB TB PB'); once rather than each call?

Johnathan Sampson's Linux example didn't work so good for me. Here's an improved version:

function getDirSize($path)
{
    $io = popen('/usr/bin/du -sb '.$path, 'r');
    $size = intval(fgets($io,80));
    pclose($io);
    return $size;
}

directory size using php filesize and RecursiveIteratorIterator.

This works with any platform which is having php 5 or higher version.

/**
 * Get the directory size
 * @param  string $directory
 * @return integer
 */
function dirSize($directory) {
    $size = 0;
    foreach(new RecursiveIteratorIterator(new RecursiveDirectoryIterator($directory)) as $file){
        $size+=$file->getSize();
    }
    return $size;
} 

function GetDirectorySize($path){
    $bytestotal = 0;
    $path = realpath($path);
    if($path!==false && $path!='' && file_exists($path)){
        foreach(new RecursiveIteratorIterator(new RecursiveDirectoryIterator($path, FilesystemIterator::SKIP_DOTS)) as $object){
            $bytestotal += $object->getSize();
        }
    }
    return $bytestotal;
}

The same idea as Janith Chinthana suggested. With a few fixes:

  • Converts $path to realpath
  • Performs iteration only if path is valid and folder exists
  • Skips . and .. files
  • Optimized for performance

There are several things you could do to optimise the script - but maximum success would make it IO-bound rather than CPU-bound:

  1. Calculate rtrim($path, '/') outside the loop.
  2. make if ($t<>"." && $t<>"..") the outer test - it doesn't need to stat the path
  3. Calculate rtrim($path, '/') . '/' . $t once per loop - inside 2) and taking 1) into account.
  4. Calculate explode(' ','B KB MB GB TB PB'); once rather than each call?

Evolved from Nate Haugs answer I created a short function for my project:

function uf_getDirSize($dir, $unit = 'm')
{
    $dir = trim($dir, '/');
    if (!is_dir($dir)) {
        trigger_error("{$dir} not a folder/dir/path.", E_USER_WARNING);
        return false;
    }
    if (!function_exists('exec')) {
        trigger_error('The function exec() is not available.', E_USER_WARNING);
        return false;
    }
    $output = exec('du -sb ' . $dir);
    $filesize = (int) trim(str_replace($dir, '', $output));
    switch ($unit) {
        case 'g': $filesize = number_format($filesize / 1073741824, 3); break;  // giga
        case 'm': $filesize = number_format($filesize / 1048576, 1);    break;  // mega
        case 'k': $filesize = number_format($filesize / 1024, 0);       break;  // kilo
        case 'b': $filesize = number_format($filesize, 0);              break;  // byte
    }
    return ($filesize + 0);
}

A pure php example.

<?php
    $units = explode(' ', 'B KB MB GB TB PB');
    $SIZE_LIMIT = 5368709120; // 5 GB
    $disk_used = foldersize("/webData/users/[email protected]");

    $disk_remaining = $SIZE_LIMIT - $disk_used;

    echo("<html><body>");
    echo('diskspace used: ' . format_size($disk_used) . '<br>');
    echo( 'diskspace left: ' . format_size($disk_remaining) . '<br><hr>');
    echo("</body></html>");


function foldersize($path) {
    $total_size = 0;
    $files = scandir($path);
    $cleanPath = rtrim($path, '/'). '/';

    foreach($files as $t) {
        if ($t<>"." && $t<>"..") {
            $currentFile = $cleanPath . $t;
            if (is_dir($currentFile)) {
                $size = foldersize($currentFile);
                $total_size += $size;
            }
            else {
                $size = filesize($currentFile);
                $total_size += $size;
            }
        }   
    }

    return $total_size;
}


function format_size($size) {
    global $units;

    $mod = 1024;

    for ($i = 0; $size > $mod; $i++) {
        $size /= $mod;
    }

    $endIndex = strpos($size, ".")+3;

    return substr( $size, 0, $endIndex).' '.$units[$i];
}

?>

Regarding Johnathan Sampson's Linux example, watch out when you are doing an intval on the outcome of the "du" function, if the size is >2GB, it will keep showing 2GB.

Replace:

$totalSize = intval(fgets($io, 80));

by:

strtok(fgets($io, 80), " ");

supposed your "du" function returns the size separated with space followed by the directory/file name.


function GetDirectorySize($path){
    $bytestotal = 0;
    $path = realpath($path);
    if($path!==false && $path!='' && file_exists($path)){
        foreach(new RecursiveIteratorIterator(new RecursiveDirectoryIterator($path, FilesystemIterator::SKIP_DOTS)) as $object){
            $bytestotal += $object->getSize();
        }
    }
    return $bytestotal;
}

The same idea as Janith Chinthana suggested. With a few fixes:

  • Converts $path to realpath
  • Performs iteration only if path is valid and folder exists
  • Skips . and .. files
  • Optimized for performance

directory size using php filesize and RecursiveIteratorIterator.

This works with any platform which is having php 5 or higher version.

/**
 * Get the directory size
 * @param  string $directory
 * @return integer
 */
function dirSize($directory) {
    $size = 0;
    foreach(new RecursiveIteratorIterator(new RecursiveDirectoryIterator($directory)) as $file){
        $size+=$file->getSize();
    }
    return $size;
} 

A one-liner solution. Result in bytes.

$size=array_sum(array_map('filesize', glob("{$dir}/*.*")));

Added bonus: you can simply change the file mask to whatever you like, and count only certain files (eg by extension).


if you are hosted on Linux:

passthru('du -h -s ' . $DIRECTORY_PATH)

It's better than foreach


Code adjusted to access main directory and all sub folders within it. This would return the full directory size.

function get_dir_size($directory){
    $size = 0;
    $files= glob($directory.'/*');
    foreach($files as $path){
        is_file($path) && $size += filesize($path);

        if (is_dir($path))
        {
            $size += get_dir_size($path);
        }
    }
    return $size;
}

A one-liner solution. Result in bytes.

$size=array_sum(array_map('filesize', glob("{$dir}/*.*")));

Added bonus: you can simply change the file mask to whatever you like, and count only certain files (eg by extension).


It works perfectly fine .

     public static function folderSize($dir)
     {
        $size = 0;

        foreach (glob(rtrim($dir, '/') . '/*', GLOB_NOSORT) as $each) {
            $func_name = __FUNCTION__;
            $size += is_file($each) ? filesize($each) : static::$func_name($each);
        }

        return $size;
      }

Thanks to Jonathan Sampson, Adam Pierce and Janith Chinthana I did this one checking for most performant way to get the directory size. Should work on Windows and Linux Hosts.

static function getTotalSize($dir)
{
    $dir = rtrim(str_replace('\\', '/', $dir), '/');

    if (is_dir($dir) === true) {
        $totalSize = 0;
        $os        = strtoupper(substr(PHP_OS, 0, 3));
        // If on a Unix Host (Linux, Mac OS)
        if ($os !== 'WIN') {
            $io = popen('/usr/bin/du -sb ' . $dir, 'r');
            if ($io !== false) {
                $totalSize = intval(fgets($io, 80));
                pclose($io);
                return $totalSize;
            }
        }
        // If on a Windows Host (WIN32, WINNT, Windows)
        if ($os === 'WIN' && extension_loaded('com_dotnet')) {
            $obj = new \COM('scripting.filesystemobject');
            if (is_object($obj)) {
                $ref       = $obj->getfolder($dir);
                $totalSize = $ref->size;
                $obj       = null;
                return $totalSize;
            }
        }
        // If System calls did't work, use slower PHP 5
        $files = new \RecursiveIteratorIterator(new \RecursiveDirectoryIterator($dir));
        foreach ($files as $file) {
            $totalSize += $file->getSize();
        }
        return $totalSize;
    } else if (is_file($dir) === true) {
        return filesize($dir);
    }
}

Just another function using native php functions.

function dirSize($dir)
    {
        $dirSize = 0;
        if(!is_dir($dir)){return false;};
        $files = scandir($dir);if(!$files){return false;}
        $files = array_diff($files, array('.','..'));

        foreach ($files as $file) {
            if(is_dir("$dir/$file")){
                 $dirSize += dirSize("$dir/$file");
            }else{
                $dirSize += filesize("$dir/$file");
            }
        }
        return $dirSize;
    }

NOTE: this function returns the files sizes, NOT the size on disk


There are several things you could do to optimise the script - but maximum success would make it IO-bound rather than CPU-bound:

  1. Calculate rtrim($path, '/') outside the loop.
  2. make if ($t<>"." && $t<>"..") the outer test - it doesn't need to stat the path
  3. Calculate rtrim($path, '/') . '/' . $t once per loop - inside 2) and taking 1) into account.
  4. Calculate explode(' ','B KB MB GB TB PB'); once rather than each call?

This is the simplest possible algorithm to find out directory size irrespective of the programming language you are using. For PHP specific implementation. go to: Calculate Directory Size in PHP | Explained with Algorithm | Working Code


A pure php example.

<?php
    $units = explode(' ', 'B KB MB GB TB PB');
    $SIZE_LIMIT = 5368709120; // 5 GB
    $disk_used = foldersize("/webData/users/[email protected]");

    $disk_remaining = $SIZE_LIMIT - $disk_used;

    echo("<html><body>");
    echo('diskspace used: ' . format_size($disk_used) . '<br>');
    echo( 'diskspace left: ' . format_size($disk_remaining) . '<br><hr>');
    echo("</body></html>");


function foldersize($path) {
    $total_size = 0;
    $files = scandir($path);
    $cleanPath = rtrim($path, '/'). '/';

    foreach($files as $t) {
        if ($t<>"." && $t<>"..") {
            $currentFile = $cleanPath . $t;
            if (is_dir($currentFile)) {
                $size = foldersize($currentFile);
                $total_size += $size;
            }
            else {
                $size = filesize($currentFile);
                $total_size += $size;
            }
        }   
    }

    return $total_size;
}


function format_size($size) {
    global $units;

    $mod = 1024;

    for ($i = 0; $size > $mod; $i++) {
        $size /= $mod;
    }

    $endIndex = strpos($size, ".")+3;

    return substr( $size, 0, $endIndex).' '.$units[$i];
}

?>

Code adjusted to access main directory and all sub folders within it. This would return the full directory size.

function get_dir_size($directory){
    $size = 0;
    $files= glob($directory.'/*');
    foreach($files as $path){
        is_file($path) && $size += filesize($path);

        if (is_dir($path))
        {
            $size += get_dir_size($path);
        }
    }
    return $size;
}

Johnathan Sampson's Linux example didn't work so good for me. Here's an improved version:

function getDirSize($path)
{
    $io = popen('/usr/bin/du -sb '.$path, 'r');
    $size = intval(fgets($io,80));
    pclose($io);
    return $size;
}

Thanks to Jonathan Sampson, Adam Pierce and Janith Chinthana I did this one checking for most performant way to get the directory size. Should work on Windows and Linux Hosts.

static function getTotalSize($dir)
{
    $dir = rtrim(str_replace('\\', '/', $dir), '/');

    if (is_dir($dir) === true) {
        $totalSize = 0;
        $os        = strtoupper(substr(PHP_OS, 0, 3));
        // If on a Unix Host (Linux, Mac OS)
        if ($os !== 'WIN') {
            $io = popen('/usr/bin/du -sb ' . $dir, 'r');
            if ($io !== false) {
                $totalSize = intval(fgets($io, 80));
                pclose($io);
                return $totalSize;
            }
        }
        // If on a Windows Host (WIN32, WINNT, Windows)
        if ($os === 'WIN' && extension_loaded('com_dotnet')) {
            $obj = new \COM('scripting.filesystemobject');
            if (is_object($obj)) {
                $ref       = $obj->getfolder($dir);
                $totalSize = $ref->size;
                $obj       = null;
                return $totalSize;
            }
        }
        // If System calls did't work, use slower PHP 5
        $files = new \RecursiveIteratorIterator(new \RecursiveDirectoryIterator($dir));
        foreach ($files as $file) {
            $totalSize += $file->getSize();
        }
        return $totalSize;
    } else if (is_file($dir) === true) {
        return filesize($dir);
    }
}

Object Oriented Approach :

/**
 * Returns a directory size
 *
 * @param string $directory
 *
 * @return int $size directory size in bytes
 *
 */
function dir_size($directory)
{
    $size = 0;
    foreach(new RecursiveIteratorIterator(new RecursiveDirectoryIterator($directory)) as $file)
    {
        $size += $file->getSize();
    }
    return $size;
}

Fast and Furious Approach :

function dir_size2($dir)
{
    $line = exec('du -sh ' . $dir);
    $line = trim(str_replace($dir, '', $line));
    return $line;
}

This is the simplest possible algorithm to find out directory size irrespective of the programming language you are using. For PHP specific implementation. go to: Calculate Directory Size in PHP | Explained with Algorithm | Working Code