unpack()

(PHP 4, PHP 5, PHP 7)

Unpack data from binary string

说明

unpack(string $format,string $data[,int $offset= 0]): array

Unpacks from a binary string into an array according to the given$format.

The unpacked data is stored in an associative array. To accomplish this you have to name the different format codes and separate them by a slash /. If a repeater argument is present, then each of the array keys will have a sequence number behind the given name.

参数

$format

Seepack()for an explanation of the format codes.

$data

The packed data.

$offset

The offset to begin unpacking from.

返回值

Returns an associative array containing unpacked elements of binary string.

更新日志

范例

Example #1unpack()example

<?php
$binarydata = "\x04\x00\xa0\x00";
$array = unpack("cchars/nint", $binarydata);
print_r($array);
?>

以上例程会输出：

Array
(
    [chars] => 4
    [int] => 160
)

Example #2unpack()example with a repeater

<?php
$binarydata = "\x04\x00\xa0\x00";
$array = unpack("c2chars/nint", $binarydata);
print_r($array);
?>

以上例程会输出：

Array
(
    [chars1] => 4
    [chars2] => 0
    [int] => 40960
)

注释

Note that PHP internally stores integral values as signed. If you unpack a large unsigned long and it is of the same size as PHP internally stored values the result will be a negative number even though unsigned unpacking was specified.

If you do not name an element, numeric indices starting from1are used. Be aware that if you have more than one unnamed element, some data is overwritten because the numbering restarts from1for each element.

Example #3unpack()example with unnamed keys

<?php
$binarydata = "\x32\x42\x00\xa0";
$array = unpack("c2/n", $binarydata);
var_dump($array);
?>

以上例程会输出：

array(2) {
  [1]=>
  int(160)
  [2]=>
  int(66)
}

Note that the first value from thecspecifier is overwritten by the first value from thenspecifier.

参见

• pack() 将数据打包成二进制字符串

A helper class to convert integer to binary strings and vice versa. Useful for writing and reading integers to / from files or sockets.
<?php
  class int_helper
  {
    public static function int8($i) {
      return is_int($i) ? pack("c", $i) : unpack("c", $i)[1];
    }
    public static function uInt8($i) {
      return is_int($i) ? pack("C", $i) : unpack("C", $i)[1];
    }
    public static function int16($i) {
      return is_int($i) ? pack("s", $i) : unpack("s", $i)[1];
    }
    public static function uInt16($i, $endianness=false) {
      $f = is_int($i) ? "pack" : "unpack";
      if ($endianness === true) { // big-endian
        $i = $f("n", $i);
      }
      else if ($endianness === false) { // little-endian
        $i = $f("v", $i);
      }
      else if ($endianness === null) { // machine byte order
        $i = $f("S", $i);
      }
      return is_array($i) ? $i[1] : $i;
    }
    public static function int32($i) {
      return is_int($i) ? pack("l", $i) : unpack("l", $i)[1];
    }
    public static function uInt32($i, $endianness=false) {
      $f = is_int($i) ? "pack" : "unpack";
      if ($endianness === true) { // big-endian
        $i = $f("N", $i);
      }
      else if ($endianness === false) { // little-endian
        $i = $f("V", $i);
      }
      else if ($endianness === null) { // machine byte order
        $i = $f("L", $i);
      }
      return is_array($i) ? $i[1] : $i;
    }
    public static function int64($i) {
      return is_int($i) ? pack("q", $i) : unpack("q", $i)[1];
    }
    public static function uInt64($i, $endianness=false) {
      $f = is_int($i) ? "pack" : "unpack";
      if ($endianness === true) { // big-endian
        $i = $f("J", $i);
      }
      else if ($endianness === false) { // little-endian
        $i = $f("P", $i);
      }
      else if ($endianness === null) { // machine byte order
        $i = $f("Q", $i);
      }
      return is_array($i) ? $i[1] : $i;
    }
  }
?>
Usage example:
<?php
  Header("Content-Type: text/plain");
  include("int_helper.php");
  echo int_helper::uInt8(0x6b) . PHP_EOL; // k
  echo int_helper::uInt8(107) . PHP_EOL; // k
  echo int_helper::uInt8("\x6b") . PHP_EOL . PHP_EOL; // 107
  echo int_helper::uInt16(4101) . PHP_EOL; // \x05\x10
  echo int_helper::uInt16("\x05\x10") . PHP_EOL; // 4101
  echo int_helper::uInt16("\x05\x10", true) . PHP_EOL . PHP_EOL; // 1296
  echo int_helper::uInt32(2147483647) . PHP_EOL; // \xff\xff\xff\x7f
  echo int_helper::uInt32("\xff\xff\xff\x7f") . PHP_EOL . PHP_EOL; // 2147483647
  // Note: Test this with 64-bit build of PHP
  echo int_helper::uInt64(9223372036854775807) . PHP_EOL; // \xff\xff\xff\xff\xff\xff\xff\x7f
  echo int_helper::uInt64("\xff\xff\xff\xff\xff\xff\xff\x7f") . PHP_EOL . PHP_EOL; // 9223372036854775807
?>

I had a situation where I had to unpack a file filled with little-endian order double-floats in a way that would work on either little-endian or big-endian machines. PHP doesn't have a formatting code that will change the byte order of doubles, so I wrote this workaround.
<?php
/*The following code is a workaround for php's unpack function
which does not have the capability of unpacking double precision
floats that were packed in the opposite byte order of the current
machine.
*/
function big_endian_unpack ($format, $data) {
  $ar = unpack ($format, $data);
  $vals = array_values ($ar);
  $f = explode ('/', $format);
  $i = 0;
  foreach ($f as $f_k => $f_v) {
  $repeater = intval (substr ($f_v, 1));
  if ($repeater == 0) $repeater = 1;
  if ($f_v{1} == '*')
  {
    $repeater = count ($ar) - $i;
  }
  if ($f_v{0} != 'd') { $i += $repeater; continue; }
  $j = $i + $repeater;
  for ($a = $i; $a < $j; ++$a)
  {
    $p = pack ('d',$vals[$i]);
    $p = strrev ($p);
    list ($vals[$i]) = array_values (unpack ('d1d', $p));
    ++$i;
  }
  }
  $a = 0;
  foreach ($ar as $ar_k => $ar_v) {
  $ar[$ar_k] = $vals[$a];
  ++$a;
  }
  return $ar;
}
list ($endiantest) = array_values (unpack ('L1L', pack ('V',1)));
if ($endiantest != 1) define ('BIG_ENDIAN_MACHINE',1);
if (defined ('BIG_ENDIAN_MACHINE')) $unpack_workaround = 'big_endian_unpack';
else $unpack_workaround = 'unpack';
?>
This workaround is used like this:
<?php
function foo() {
    global $unpack_workaround;
  $bar = $unpack_workaround('N7N/V2V/d8d',$my_data);
//...
}
?>
On a little endian machine, $unpack_workaround will simply point to the function unpack. On a big endian machine, it will call the workaround function.
Note, this solution only works for doubles. In my project I had no need to check for single precision floats.

This is about the last example of my previous post. For the sake of clarity, I'm including again here the example, which expands the one given in the formal documentation:
<?
 $binarydata = "AA\0A";
 $array = unpack("c2chars/nint", $binarydata);
 foreach ($array as $key => $value)
   echo "\$array[$key] = $value <br>\n";
?>
This outputs:
$array[chars1] = 65 
$array[chars2] = 65 
$array[int] = 65 
Here, we assume that the ascii code for character 'A' is decimal 65.
Remebering that the format string structure is:
<format-code> [<count>] [<array-key>] [/ ...],
in this example, the format string instructs the function to
 1. ("c2...") Read two chars from the second argument ("AA ...), 
 2. (...chars...) Use the array-keys "chars1", and "chars2" for 
   these two chars read,
 3. (.../n...) Read a short int from the second argument (...\0A"),
 4. (...int") Use the word "int" as the array key for the just read
   short.
I hope this is clearer now,
Sergio.

If having a zero-based index is useful/necessary, then instead of:
$int_list = unpack("s*", $some_binary_data);
 try:
$int_list = array_merge(unpack("s*", $some_binary_data));
This will return a 0-based array:
$int_list[0] = x
$int_list[1] = y
$int_list[2] = z
...
rather than the default 1-based array returned from unpack when no key is supplied:
$int_list[1] = x
$int_list[2] = y
$int_list[3] = z
...
It's not used often, but array_merge() with only one parameter will compress a sequentially-ordered numeric-index, starting with an index of [0].

Suppose we need to get some kind of internal representation of an integer, say 65, as a four-byte long. Then we use, something like:
<?
 $i = 65;
 $s = pack("l", $i); // long 32 bit, machine byte order
 echo strlen($s) . "<br>\n";
 echo "***$s***<br>\n";
?>
The output is:
X-Powered-By: PHP/4.1.2
Content-type: text/html
4
***A*** 
(That is the string "A\0\0\0")
Now we want to go back from string "A\0\0\0" to number 65. In this case we can use:
<?
 $s = "A\0\0\0"; // This string is the bytes representation of number 65
 $arr = unpack("l", $s);
 foreach ($arr as $key => $value)
   echo "\$arr[$key] = $value<br>\n";
?>
And this outpus:
X-Powered-By: PHP/4.1.2
Content-type: text/html
$arr[] = 65
Let's give the array key a name, say "mykey". In this case, we can use: 
<?
 $s = "A\0\0\0"; // This string is the bytes representation of number 65
 $arr = unpack("lmykey", $s);
 foreach ($arr as $key => $value)
   echo "\$arr[$key] = $value\n";
?>
An this outpus:
X-Powered-By: PHP/4.1.2
Content-type: text/html
$arr[mykey] = 65
The "unpack" documentation is a little bit confusing. I think a more complete example could be:
<?
 $binarydata = "AA\0A";
 $array = unpack("c2chars/nint", $binarydata);
 foreach ($array as $key => $value)
  echo "\$array[$key] = $value <br>\n";
?>
whose output is:
X-Powered-By: PHP/4.1.2
Content-type: text/html
$array[chars1] = 65 <br>
$array[chars2] = 65 <br>
$array[int] = 65 <br>
Note that the format string is something like
<format-code> [<count>] [<array-key>] [/ ...]
I hope this clarifies something
Sergio

be aware of the behavior of your system that PHP resides on.
On x86, unpack MAY not yield the result you expect for UInt32
This is due to the internal nature of PHP, being that integers are internally stored as SIGNED!
For x86 systems, unpack('N', "\xff\xff\xff\xff") results in -1
For (most?) x64 systems, unpack('N', "\xff\xff\xff\xff") results in 4294967295.
This can be verified by checking the value of PHP_INT_SIZE.
If this value is 4, you have a PHP that internally stores 32-bit.
A value of 8 internally stores 64-bit.
To work around this 'problem', you can use the following code to avoid problems with unpack.
The code is for big endian order but can easily be adjusted for little endian order (also, similar code works for 64-bit integers):
<?php
function _uint32be($bin)
{
  // $bin is the binary 32-bit BE string that represents the integer
  if (PHP_INT_SIZE <= 4){
    list(,$h,$l) = unpack('n*', $bin);
    return ($l + ($h*0x010000));
  }
  else{
    list(,$int) = unpack('N', $bin);
    return $int;
  }
}
?>
Do note that you *could* also use sprintf('%u', $x) to show the unsigned real value.
Also note that (at least when PHP_INT_SIZE = 4) the result WILL be a float value when the input is larger then 0x7fffffff (just check with gettype);
Hope this helps people.

The documentation is clear that an integer read using an unsigned format character will still be stored as a signed integer. The often-cited work-around is to use sprintf('%u', $bigint) to properly display integers with the MSB set.
In the case where the numeric value is more important than how it's displayed, you can still work with other large integers using intval() to "upgrade" your existing unsigned integers.
I had a problem comparing 32-bit integers read from files with hard-coded constants (file signatures tend to need this). Here's what I did to avoid converting everything into strings:
<?php
$bigint = 0x89504E47;
$packed = pack('N', $bigint);
list($unpacked) = array_values(unpack('N', $packed));
//The $bigint remains an unsigned integer.
//Even though their bit-wise values are identical, comparison fails.
echo 'bigint ',
 ($bigint == $unpacked ? '==' : '!='),
 " unpacked\n";
//intval() triggers a re-interpretation of $bigint.
//$bigint is internally compared as a signed integer.
//Since the bit-wise value of $bigint never changes, comparison succeeds.
echo 'intval(bigint) ',
 (intval($bigint) == $unpacked ? '==' : '!='),
 " unpacked\n";
?>
It works, but it's a little backwards. If anyone has any ideas on how to "downgrade" a signed integer into an unsigned integer without using strings, that would be a valuable note to add to the documentation.

Another option for converting binary data into PHP data types, is to use the Zend Framework's Zend_Io_Reader class:
http://bit.ly/9zAhgz
There's also a Zend_Io_Writer class that does the reverse.

Reading a text cell from an Excel spreadsheet returned a string with low-order embedded nulls: 0x4100 0x4200 etc. To remove the nulls, used
<?php
$strWithoutNulls = implode( '', explode( "\0", $strWithNulls ) );
?>
(unpack() didn't seem to help much here; needed chars back to re-constitute the string, not integers.)

If no key name is given [e.g., unpack('C*',$data)], the keys are simply integers starting at 1, and you have a standard array. (I know of no way to get the array to start at zero.)
If you use multiple types, you must give a key name for all of them (except optionally one), because the key counter is reset with each slash. For example, in unpack('n2/C*',$data), indices 1 and 2 of the returned array are filled by integers ('n'), then overwritten with characters ('C').

As stated above, "if you unpack a large unsigned long and it is of the same size as PHP internally stored values the result will be a negative number even though unsigned unpacking was specified."
To restore the original unsigned value, you could do this :
if ($unpackedVal <0)
{
   $unpackedVal += 4294967296;
}
Hope this helps !
Cheers

Warning: This unpack function makes the array with keys starting at 1 instead of starting at 0.
For example:
<?php
 function read_field($h) {
 $a=unpack("V",fread($h,4));
 return fread($h,$a[1]);
 }
?>

If you just want to extract a dword/long int from a binary string, the following code works beautifully (intel endian):
$Number = ord($Buffer{0}) | (ord($Buffer{1})<<8) | (ord($Buffer{2})<<16) | (ord($Buffer{3})<<24);

Here's a demonstration concerning the speed of unpacking files:
So let's see which method is fastest between FREAD or SUBSTR?
I was creating a script that could read scenario files from a game, and render a preview of its terrain. The terrain structure within each file was huge (between 100,000 - 1,000,000 blocks containing 3 bits of data each). Therefore, I spent much effort to ensure it was fast and robust.
Method 1: This method retrieves the 3 bits of data found in each block. It uses the loop of widthxheight and implode+unpack+substr each block:
<?php
for ( $Y = 0 ; $Y < ( $width * $height ) ; $Y ++ ) {
  $Output [ Map ] [ $Y ] [ TerrainID ] = implode ( null , unpack ( 'c1' , substr ( $Input , $Line ) ) ) ; $Line += 1 ;
  $Output [ Map ] [ $Y ] [ Elevation ] = implode ( null , unpack ( 'c1' , substr ( $Input , $Line ) ) ) ; $Line += 1 ;
  $Output [ Map ] [ $Y ] [ Unknown ] = implode ( null , unpack ( 'c1' , substr ( $Input , $Line ) ) ) ; $Line += 1 ;
}
//The average microtime was: 2.9 sec
?>
Note that it takes even more time if you use a custom function to implement the implode+unpack+substr functions.
Now... This method uses the FREAD function:
<?php
for ( $Y = 0 ; $Y < ( $width * $height ) ; $Y ++ ) {
  $Output [ Map ] [ $Y ] = unpack ( 'c3' , fread ( $sc , 3 ) ) ;
}
//Average microtime was: 0.7 sec
?>
I recommend using the FREAD method instead of SUBSTR.
Another test!!! This method is 10x faster than the above. This does not use the FOR loop:
<?php
$Output [ Map ] [ Data ] = unpack ( 'c' . ( $width * $height ) , stream_get_contents ( $sc ) ) ;
//Average microtime: 0.08 - 0.05 sec
?>
If you want to read files much faster, you should try to reduce the number of loops and use the unpack function to its simplest and robust method.

The script following is a example how to save more than one values on file separating its with "\r\n" and how to recovering its values.
<?php
// Save two integer values in a binary file
$nomearq = "./teste.bin";
$valor = 123;
$ptrarq = fopen($nomearq, "wb");
$valorBin = pack("L",$valor);
echo "First value ($valor) packed with ";
echo fwrite($ptrarq, $valorBin)." bytes<br>";
echo "Separator \\r\\n with ";
echo fwrite($ptrarq, "\r\n")." bytes<br>";
$valor = 456;
$valorBin = pack("L",$valor);
echo "Second value ($valor) packed with ";
echo fwrite($ptrarq, $valorBin)." bytes<br>";
fclose($ptrarq);
// Recover the saved values
$ptrarq = fopen($nomearq, "rb");
$valorBin = file($nomearq,filesize($nomearq));
echo "<br>The reading values is:<br>";
foreach($valorBin as $valor){
 $valor = unpack("L",$valor);
 print_r($valor);
 echo "<br>";
}
fclose($ptrarq);
?>
Results:
First value (123) packed with 4 bytes
Separator \r\n with 2 bytes
Second value (456) packed with 4 bytes
The reading values is:
Array ( [1] => 123 ) 
Array ( [1] => 456 )

As with perl, the count for hex is number of nybbles or half-bytes, this differs from the other options which count in full bytes.

A simpler solution is to mask the value with 0xffffffff. For instance:
<?php
$rec = unpack(
 "Vvalue/".
 "Vhash32/",
 $recbin);
$rec['hash32'] &= 0xffffffff;
$rec['value'] &= 0xffffffff;
?>
Unlike sprintf(), which converts the value to a string, this preserves the numeric type of the value.

Functions I found useful when dealing with fixed width file processing, related to unpack/pack functions.
<?php
/**
* funpack
* format: array of key, length pairs
* data: string to unpack
*/
function funpack($format, $data){
  foreach ($format as $key => $len) {
    $result[$key] = trim(substr($data, $pos, $len));
    $pos+= $len;
  }
  return $result;
}
/**
* fpack
* format: array of key, length pairs
* data: array of key, value pairs to pack
* pad: padding direction
*/
function fpack($format, $data, $pad = STR_PAD_RIGHT){
  foreach ($format as $key => $len){
    $result .= substr(str_pad($data[$key], $len, $pad), 0, $len);
  }
  return $result;
}
?>

Above it says this:
 "Note that PHP internally stores integral values as signed. If you unpack a large unsigned long and it is of the same size as PHP internally stored values the result will be a negative number even though unsigned unpacking was specified."
This happened to me. I wanted to get a big number from a unsigned long, but it kept coming returning a negative. Happened to notice that sprintf('%u',$dta) will take the useless negative and restore it into its large unsigned proper magnitude.
Hope this saves someone a little time...

To convert big endian to little endian or to convert little endian to big endian, use the following approach as an example:
<?php
// file_get_contents() returns a binary value, unpack("V*", _ ) returns an unsigned long 32-bit little endian decimal value, but bin2hex() after that would just give the hex data in the file if alone, so instead we use:
// file_get_contents(), unpack("V*", _ ), then dechex(), in that order, to get the byte-swapping effect.
?>
With the logic of the approach in this example, you can discover how to swap the endian byte order as you need.

I wrote a quick pair of functions using pack/unpack for converting between raw binary (e.g. openssl_random_pseudo_bytes() output) and hexadecimal (e.g. hash() output):
<?php
function raw2hex($raw) {
 $m = unpack('H*', $raw);
 return $m[1];
}
function hex2raw($hex) { 
 return pack('H*', $hex);
}
?>
Feel free to suggest any improvements, but I thought this was worth sharing.

I hadn't realized that if the number after the unpack type was 1 (i.e. "V1page"), that it would behave as if there was no number at all. I had been using a variable and didn't think to watch for this. For instance,
<?php
if ($something)
  $get = 2;
else
  $get = 1;
$arr = unpack("V" . $get . "page", $data);
?>
Now if $something was FALSE, then $arr will only have one entry named "page". If $something was TRUE, $arr would have "page1" and "page2".

<?php
function parse_pascalstr($bytes_parsed, $parse_str) { 
  $parse_info = unpack("x$bytes_parsed/cstr_len", $parse_str);
  $str_len = $parse_info["str_len"]; 
  $bytes_parsed = $bytes_parsed + 1; 
  $parse_info = unpack("x$bytes_parsed/A".$str_len."str", $parse_str); 
  $str = $parse_info["str"]; 
  $bytes_parsed = $bytes_parsed + strlen($str); 
  return array($str, $bytes_parsed); 
}
?>

You might find these functions useful:
<?php
function byteStr2byteArray($s) {
    return array_slice(unpack("C*", "\0".$s), 1);
}
function byteArray2byteStr(array $t) {
    return call_user_func_array(pack, array_merge(array("C*"), $t));
}
function lsbStr2ushortArray($s) {
    return array_slice(unpack("v*", "\0\0".$s), 1);
}
function ushortArray2lsbStr(array $t) {
    return call_user_func_array(pack, array_merge(array("v*"), $t));
}
function lsbStr2ulongArray($s) {
    return array_slice(unpack("V*", "\0\0\0\0".$s), 1);
}
function ulongArray2lsbStr(array $t) {
    return call_user_func_array(pack, array_merge(array("V*"), $t));
}
?>
Of course, you can address byte strings as if they're arrays with numerical indexes, but the other functions are helpful.

Here is my solution to reading a Big-Endian formatted double on an Little-Endian machine.
<?php
function ToDouble($data) {
  $t = unpack("C*", pack("S*", 256));
  if($t[1] == 1) {
    $a = unpack("d*", $data);
  } else {
    $a = unpack("d*", strrev($data));
  }
  return (double)$a[1];
}
?>