gzopen

(PHP 4, PHP 5, PHP 7)

gzopen — Open gz-file

Description

resource gzopen ( string $filename , string $mode [, int $use_include_path = 0 ] )

Opens a gzip (.gz) file for reading or writing.

gzopen() can be used to read a file which is not in gzip format; in this case gzread() will directly read from the file without decompression.

Parameters

filename: The file name.
mode: As in fopen() (rb or wb) but can also include a compression level (wb9) or a strategy: f for filtered data as in wb6f, h for Huffman only compression as in wb1h. (See the description of deflateInit2 in zlib.h for more information about the strategy parameter.)
use_include_path: You can set this optional parameter to 1, if you want to search for the file in the include_path too.

Return Values

Returns a file pointer to the file opened, after that, everything you read from this file descriptor will be transparently decompressed and what you write gets compressed.

If the open fails, the function returns FALSE.

Examples

Example #1 gzopen() Example


<?php
$fp = gzopen("/tmp/file.gz", "r");
?>

User Contributed Notes

David Gero dave at havidave dot com

4 years ago


gzopen("php://output","wb") doesn't work on a web server, nor does fopen("compress.zlib://php://output","wb").

Here is a snippet to gzip a file and output it on the fly, without using a temporary file, without reading the file into memory, and without reading the file more than once:

<?php
$fin = fopen($file, "rb");
if ($fin !== FALSE) {
    $fout = fopen("php://output", "wb");
    if ($fout !== FALSE) {
    // write gzip header
    fwrite($fout, "\x1F\x8B\x08\x08".pack("V", filemtime($file))."\0\xFF", 10);
    // write the original file name
    $oname = str_replace("\0", "", basename($file));
    fwrite($fout, $oname."\0", 1+strlen($oname));
    // add the deflate filter using default compression level
    $fltr = stream_filter_append($fout, "zlib.deflate", STREAM_FILTER_WRITE, -1);
    // set up the CRC32 hashing context
    $hctx = hash_init("crc32b");
    // turn off the time limit
    if (!ini_get("safe_mode")) set_time_limit(0);
    $con = TRUE;
    $fsize = 0;
    while (($con !== FALSE) && !feof($fin)) {
        // deflate works best with buffers >32K
        $con = fread($fin, 64 * 1024);
        if ($con !== FALSE) {
        hash_update($hctx, $con);
        $clen = strlen($con);
        $fsize += $clen;
        fwrite($fout, $con, $clen);
        }
    }
    // remove the deflate filter
    stream_filter_remove($fltr);
    // write the CRC32 value
    // hash_final is a string, not an integer
    $crc = hash_final($hctx, TRUE);
    // need to reverse the hash_final string so it's little endian
    fwrite($fout, $crc[3].$crc[2].$crc[1].$crc[0], 4);
    // write the original uncompressed file size
    fwrite($fout, pack("V", $fsize), 4);
    fclose($fout);
    }
    fclose($fin);
}
?>

dj8 at joesvolcano dot net

7 months ago


An OO-version of David Gero's excellent GZ File format note also on this page:

<?php

// David Gero's example: read a file and output GZ
$gz_to_out = new GZTempFile(basename($file), "php://output");
$fin = fopen($file, "rb");
while($data = fread($fin, 64 * 1024)) {
    $gz_to_out->fwrite();
}
close($fin);
$gz_to_out->flushBuffer();

// Example of building your GZ file content on the fly (temp filehandle)
$gz_custom = new GZTempFile();
foreach ( ... ) {
    //  Some work
    $str = ...

    $gz_custom->fwrite($str);
}
//  Store it in a database
$m = new MongoClient();
$gridfs = $m->selectDB('test')->getGridFS();
$id = $gridfs->storeFile($gz_custom->getReadFilehandle(), array('contentType' => 'application/x-gzip'));

class GZTempFile {
    private $__fh = null;
    public $uncompressed_bytes = 0;
    public $filesize = null;
    private $gz_filter = null;
    private $file_hash = null;
    private $final_read_fh = false;
    private $__buffer = '';
    private $__buffer_len = 0;

    public function __construct($filename = 'data', $fh = null) {
        $this->__fh = is_null($fh) ? fopen('php://temp','w+') : $fh;
        fwrite($this->__fh, "\x1F\x8B\x08\x08".pack("V", time())."\0\xFF", 10); // GZ file header
        fwrite($this->__fh, str_replace("\0", "", basename($filename)) ."\0");  // GZ filename = data, needed???
        $this->gz_filter = stream_filter_append($this->__fh, "zlib.deflate", STREAM_FILTER_WRITE, -1);
        $this->uncompressed_bytes = 0;
        $this->file_hash = hash_init("crc32b");
    }

    public function fwrite($str,$length = null) {
        if ( $this->final_read_fh ) { throw new Exception("GZTempFile has already been finalized and closed.  No more writing"); }
        hash_update($this->file_hash, $str);
        $this->uncompressed_bytes += strlen($str);
        $this->__buffer_len += strlen($str);
        $this->__buffer .= $str;
        if ( $this->__buffer_len >= 64 * 1024 ) { $this->flushBuffer(); }
    }
    public function flushBuffer() {
        if ( $this->__buffer_len == 0 ) { return false; }
        $return = fwrite($this->__fh, $this->__buffer);
        $this->__buffer_len = 0;
        $this->__buffer = '';
        return $return;
    }

    public function getReadFilehandle() {        
        if ( ! $this->final_read_fh ) {
            $this->flushBuffer();
            stream_filter_remove($this->gz_filter);
            $crc = hash_final($this->file_hash, TRUE);            // hash_final is a string, not an integer
            fwrite($this->__fh, $crc[3].$crc[2].$crc[1].$crc[0]); // need to reverse the hash_final string so it's little endian
            fwrite($this->__fh, pack("V", $this->uncompressed_bytes), 4);

            $this->filesize = ftell($this->__fh);
            rewind($this->__fh);
            $this->final_read_fh = $this->__fh;
        }
        return $this->final_read_fh;
    }
}

plasma

7 years ago


This worked unstable for me under high load (50+ files per second):

<?php
    $gz = gzopen ( $file, 'w9' );
    gzwrite ( $gz, $content );
    gzclose ( $gz );
?>

The following works fine:

<?php
    $f = fopen ( $file, 'w' );
    fwrite ( $f, gzcompress ( $content, 9 ) );
    fclose ( $f );
?>

katzlbtjunk at hotmail dot com

8 years ago


WARNING gzopen and gzread have a major disadvantage. They makes NO checksum and NO length verification of the gzipped data and discard this valuable information. This should be documented here.

rob at digital-crocus dot com

10 years ago


dtorop932 at hotmail dot com's comments, according to my tests, is incorrect. That code wishes to download the entire file before parsing, which is inconvinient. The wget method works though.

pentek_imre at mailbox dot hu

11 years ago


Be aware that when opening a remote file on a http server the gzopen will return by default false after 120 seconds waiting to any answer.

dtorop932 at hotmail dot com