A better example, to illustrate the differences in speed for large files, between fgets and stream_get_line.
This example simulates situations where you are reading potentially very long lines, of an uncertain length (but with a maximum buffer size), from an input source.
As Dade pointed out, the previous example I provided was much to easy to pick apart, and did not adequately highlight the issue I was trying to address.
Note that specifying a definitive end-character for fgets (ie: newline), generally decreases the speed difference reasonably significantly.
#!/usr/bin/php
<?php
$plaintext=file_get_contents('http://loripsum.net/api/60/verylong/plaintext'); # Should be around 90k characters
$plaintext=str_replace("\n"," ",$plaintext); # Get rid of newlines
$fp=fopen("/tmp/SourceFile.txt","w");
for($i=0;$i<100000;$i++) {
fputs($fp,substr($plaintext,0,rand(4096,65534)) . "\n");
}
fclose($fp);
$fp=fopen("/tmp/SourceFile.txt","r");
$start=microtime(true);
while($line=fgets($fp,65535)) {
1;
}
$end=microtime(true);
fclose($fp);
$delta1=($end - $start);
$fp=fopen("/tmp/SourceFile.txt","r");
$start=microtime(true);
while($line=stream_get_line($fp,65535)) {
1;
}
$end=microtime(true);
fclose($fp);
$delta2=($end - $start);
$pdiff=$delta1/$delta2;
print "stream_get_line is " . ($pdiff>1?"faster":"slower") . " than fgets - pdiff is $pdiff\n";
?>
$ ./testcase.php
stream_get_line is faster than fgets - pdiff is 1.760398041785
Note that, in a vast majority of situations in which php is employed, tiny differences in speed between system calls are of negligible importance.
fgets
(PHP 4, PHP 5, PHP 7, PHP 8)
fgets — 从文件指针中读取一行
参数
-
stream -
文件指针必须是有效的，必须指向由 fopen() 或 fsockopen() 成功打开的文件(并还未由 fclose() 关闭)。
-
length -
从
handle指向的文件中读取一行并返回长度最多为length- 1 字节的字符串。碰到换行符（包括在返回值中）、EOF 或者已经读取了 length - 1 字节后停止（看先碰到那一种情况）。如果没有指定length，则默认为 1K，或者说 1024 字节。
返回值
从指针 stream 指向的文件中读取了 length - 1 字节后返回字符串。
如果文件指针中没有更多的数据了则返回 false。
错误发生时返回 false。
示例
示例 #1 逐行读取文件
<?php
$fp = @fopen("/tmp/inputfile.txt", "r");
if ($fp) {
while (($buffer = fgets($fp, 4096)) !== false) {
echo $buffer;
}
if (!feof($fp)) {
echo "Error: unexpected fgets() fail\n";
}
fclose($fp);
}
?>注释
注意: 在读取在 Macintosh 电脑中或由其创建的文件时， 如果 PHP 不能正确的识别行结束符，启用运行时配置可选项 auto_detect_line_endings 也许可以解决此问题。
注意:
习惯了 C 语言中 fgets() 语法的人应该注意到
EOF是怎样被返回的。
参见
- fgetss() - 从文件指针中读取一行并过滤掉 HTML 标记
- fread() - 读取文件（可安全用于二进制文件）
- fgetc() - 从文件指针中读取字符
- stream_get_line() - 从资源流里读取一行直到给定的定界符
- fopen() - 打开文件或者 URL
- popen() - 打开进程文件指针
- fsockopen() - 打开 Internet 或者 Unix 套接字连接
- stream_set_timeout() - Set timeout period on a stream
add a note
User Contributed Notes 38 notes
Leigh Purdie ¶
9 years ago
David at Weintraub.name ¶
17 years ago
There's an error in the documentation:
The file pointer must be valid, and must point to a file successfully opened by fopen() or fsockopen() (and not yet closed by fclose()).
You should also add "popen" and "pclose" to the documentation. I'm a new PHP developer and went to verify that I could use "fgets" on commands that I used with "popen".
Anonymous ¶
4 years ago
if you for some reason need to get lines from a string instead of a file pointer, try
<?php
function string_gets(string $source, int $offset = 0, string $delimiter = "\n"): ?string
{
$len = strlen($source);
if ($len < $offset) {
// out of bounds.. maybe i should throw an exception
return null;
}
if ($len === $offset) {
// end of string..
return null;
}
$delimiter_pos = strpos($source, $delimiter, $offset);
if ($delimiter_pos === false) {
// last line.
return substr($source, $offset);
}
return substr($source, $offset, ($delimiter_pos - $offset) + strlen($delimiter));
}
?>
(i had a ~16GB string in-memory i needed to process line-by-line, but i would get memory-allocation-crash (on a 32GB ram system) if i tried explode("\n",$str); , so came up with this.. interestingly, fgets() seems to be faster than doing it in-ram-in-php, though. php 7.3.7)
tavernadelleidee[italy] ¶
18 years ago
I think that the quickest way of read a (long) file with the rows in reverse order is
<?php
$myfile = 'myfile.txt';
$command = "tac $myfile > /tmp/myfilereversed.txt";
passthru($command);
$ic = 0;
$ic_max = 100; // stops after this number of rows
$handle = fopen("/tmp/myfilereversed.txt", "r");
while (!feof($handle) && ++$ic<=$ic_max) {
$buffer = fgets($handle, 4096);
echo $buffer."<br>";
}
fclose($handle);
?>
It echos the rows while it is reading the file so it is good for long files like logs.
Borgonovo
afwxkat at gmail dot com ¶
15 years ago
One thing I discovered with fgets, at least with PHP 5.1.6, is that you may have to use an IF statement to avoid your code running rampant (and possibly hanging the server). This can cause problems if you do not have root access on the server on which you are working.
This is the code I have implemented ($F1 is an array):
<?php
if($fh = fopen("filename","r")){
while (!feof($fh)){
$F1[] = fgets($fh,9999);
}
fclose($fh);
}
?>
I have noticed that without the IF statement, fgets seems to ignore when $fh is undefined (i.e., "filename" does not exist). If that happens, it will keep attempting to read from a nonexistent filehandle until the process can be administratively killed or the server hangs, whichever comes first.
anacreo has gmail ¶
17 years ago
I'm using this function to modify the header of a large postscript document on copy... Works extremely quickly so far...
function write($filename) {
$fh = fopen($this->sourceps,'r');
$fw = fopen($filename,'w');
while (!feof($fh)) {
$buffer = fgets($fh);
fwrite($fw,$buffer);
if (!$setupfound && ereg("^%%BeginSetup",$buffer)) {
$setupfound++;
if (array_key_exists("$filename",$this->output)) {
foreach ($this->output[$filename] as $function => $value) {
fwrite($fw,$value);
}
}
stream_copy_to_stream($fh,$fw);
}
}
fclose($fw);
fclose($fh);
}
Carlo ¶
6 years ago
Error in the example number 1 of this page.
change this line:
$buffer = fgets($fd, 4096);
into:
$buffer = fgets($handle, 4096);
Peter Schlaile ¶
17 years ago
fscanf($file, "%s\n") isn't really a good substitution for fgets(), since it will stop parsing at the first whitespace and not at the end of line!
(See the fscanf page for details on this)
d at foo.com ¶
18 years ago
For sockets, If you dont want fgets, fgetc etc... to block if theres no data there. set socket_set_blocking(handle,false); and socket_set_blocking(handle,true); to set it back again.
angelo [at] mandato <dot> com ¶
20 years ago
Sometimes the strings you want to read from a file are not separated by an end of line character. the C style getline() function solves this. Here is my version:
<?php
function getline( $fp, $delim )
{
$result = "";
while( !feof( $fp ) )
{
$tmp = fgetc( $fp );
if( $tmp == $delim )
return $result;
$result .= $tmp;
}
return $result;
}
// Example:
$fp = fopen("/path/to/file.ext", 'r');
while( !feof($fp) )
{
$str = getline($fp, '|');
// Do something with $str
}
fclose($fp);
?>
kpeters AT-AT monolithss DEE OH TEE com ¶
18 years ago
It appears that fgets() will return FALSE on EOF (before feof has a chance to read it), so this code will throw an exception:
while (!feof($fh)) {
$line = fgets($fh);
if ($line === false) {
throw new Exception("File read error");
}
}
Daniel Klein ¶
16 years ago
The file pointer that fgets() uses can also be created with the proc_open() function and used with the stdout pipe created from the executed process.
ecvej ¶
18 years ago
I would have expected the same behaviour from these bits of code:-
<?php
/*This times out correctly*/
while (!feof($fp)) {
echo fgets($fp);
}
/*This times out before eof*/
while ($line=fgets($fp)) {
echo $line;
}
/*A reasonable fix is to set a long timeout*/
stream_set_timeout($fp, 180);
while ($line=fgets($fp)) {
echo $line;
}
?>
bfb1985 at yahoo dot es ¶
8 years ago
I had loads of trouble while reading with fgets on a WAMP (Windows server). On local the file went unto a <pre> tag without a hitch, but when I moved the code to a LAMP production server, every \r\n created two fgets and I got free empty lines.
I tried deleting with $string=str_replace("\r\n","\n",$string); but it had no effect whatsoever. The solution was to do an fread() and explode the contents by PHP_EOL and do a foreach($lines as $line) so every line did not get duplicated.
Here is the example code:
$file=fopen("test.txt,"r");
$text=fread($file,filesize("test.txt"));
$lines=explode(PHP_EOL,$text);
foreach($lines as $line)
{
// Do something
}
Anonymous ¶
6 years ago
<form>
<input type='text' name='filepath' value='<?php echo ((isset($_GET["filepath"])) ? $_GET["filepath"] : "");?>'>
<br>
<select name='sel'>
<option value='var1' <?php echo ((isset($_GET["sel"]) && $_GET["sel"] == "var1") ? "selected=true" : "");?>> For Year</option>
<option value='var2' <?php echo ((isset($_GET["sel"]) && $_GET["sel"] == "var2") ? "selected=true" : "");?>>Name</option>
<option value='var3' <?php echo ((isset($_GET["sel"]) && $_GET["sel"] == "var3") ? "selected=true" : "");?>>Name&Year</option>
</select>
<br>
<input type='submit' value='Button'>
</form>
<?php
if((!isset($_GET['filepath']) || !file_exists($_GET['filepath'])) || !isset($_GET['sel']))
exit("");
echo "List<br>";
$fullPath = "D:\\OSPanel\\domains\\" . $_GET["filepath"];
$f = fopen($fullPath, "r");
$arr;
for($i = 0; $str = fgets($f); $i++){
$tempArr0 = explode('-', $str);
$arr[$i][0] = trim($tempArr0[0]);
$arr[$i][1] = trim($tempArr0[1]);
}
if($_GET["sel"] == "var1"){
sort($arr);
echo "<p>";
for($i = 0; $i < count($arr); $i++)
echo "<i>{$arr[$i][0]}</i>; ";
echo "</p>";
}
else if($_GET["sel"] == "var2"){
for($i = 0; $i < count($arr); $i++)
echo "<p><b>{$arr[$i][1]}</b></p>";
}
else if($_GET["sel"] == "var3"){
for($i = 0; $i < count($arr); $i++)
echo "<p><b>{$arr[$i][1]}</b>: <i>{$arr[$i][0]}</i></p>";
}
?>
lzsiga at freemail.c3.hu ¶
14 years ago
Some people try to call feof before fgets, and then ignoring the return value of fgets. This method leads to processing value FALSE when reaching the end of file.
Bad example:
<?php
$f = fopen ("fgetstest.php", "r");
$ln= 0;
while (! feof ($f)) {
$line= fgets ($f);
++$ln;
printf ("%2d: ", $ln);
if ($line===FALSE) print ("FALSE\n");
else print ($line);
}
fclose ($f);
?>
Good example:
<?php
$f = fopen ("fgetstest.php", "r");
$ln= 0;
while ($line= fgets ($f)) {
++$ln;
printf ("%2d: ", $ln);
if ($line===FALSE) print ("FALSE\n");
else print ($line);
}
fclose ($f);
?>
Leigh Purdie ¶
16 years ago
For large files, consider using stream_get_line rather than fgets - it can make a significant difference.
$ time yes "This is a test line" | head -1000000 | php -r '$fp=fopen("php://stdin","r"); while($line=stream_get_line($fp,65535,"\n")) { 1; } fclose($fp);'
real 0m1.482s
user 0m1.616s
sys 0m0.152s
$ time yes "This is a test line" | head -1000000 | php -r '$fp=fopen("php://stdin","r"); while($line=fgets($fp,65535)) { 1; } fclose($fp);'
real 0m7.281s
user 0m7.392s
sys 0m0.136s
dandrews OVER AT 3dohio DOT com ¶
19 years ago
Saku's example may also be used like this:
<?php
@ $pointer = fopen("$DOCUMENT_ROOT/foo.txt", "r"); // the @ suppresses errors so you have to test the pointer for existence
if ($pointer) {
while (!feof($pointer)) {
$preTEXT = fgets($pointer, 999);
// $TEXT .= $preTEXT; this is better for a string
$ATEXT[$I] = $preTEXT; // maybe better as an array
$I++;
}
fclose($pointer);
}
?>
Dade Brandon ¶
11 years ago
Regarding Leigh Purdie's comment (from 4 years ago) about stream_get_line being better for large files, I decided to test this in case it was optimized since then and I found out that Leigh's comment is just completely incorrect
fgets actually has a small amount of better performance, but the test Leigh did was not set up to produce good results
The suggested test was:
$ time yes "This is a test line" | head -1000000 | php -r '$fp=fopen("php://stdin","r"); while($line=stream_get_line($fp,65535,"\n")) { 1; } fclose($fp);'
0m1.616s
$ time yes "This is a test line" | head -1000000 | php -r '$fp=fopen("php://stdin","r"); while($line=fgets($fp,65535)) { 1; } fclose($fp);'
0m7.392s
The reason this is invalid is because the buffer size of 65535 is completely unnecessary
piping the output of "yes 'this is a test line'" in to PHP makes each line 19 characters plus the delimiter
so while I don't know why stream_get_line performs better with an oversize buffer, if both buffer sizes are correct, or default, they have a negligable performance difference - although notably, stream_get_line is consistent - however if you're thinking of switching, make sure to be aware of the difference between the two functions, that stream_get_line does NOT append the delimiter, and fgets DOES append the delimiter
Here are the results on one of my servers:
Buffer size 65535
stream_get_line: 0.340s
fgets: 2.392s
Buffer size of 1024
stream_get_line: 0m0.348s
fgets: 0.404s
Buffer size of 8192 (the default for both)
stream_get_line: 0.348s
fgets: 0.552s
Buffer size of 100:
stream_get_line: 0.332s
fgets: 0.368s
Anonymous ¶
18 years ago
Macintosh line endings mentioned in docs refer to Mac OS Classic. You don't need this setting for interoperability with unixish OS X.
timr ¶
20 years ago
If you need to read an entire file into a string, use file_get_contents(). fgets() is most useful when you need to process the lines of a file separately.
sam dot bryan at montal dot com ¶
18 years ago
An easy way to authenticate Windows Domain users from scripts running on a non-Windows or non-Domain box - pass the submitted username and password to an IMAP service on a Windows machine.
<?php
$server = 'imapserver';
$user = 'user';
$pass = 'pass';
if (authIMAP($user, $pass, $server)) {
echo "yay";
} else {
echo "nay";
}
function authIMAP($user, $pass, $server) {
$connection = fsockopen($server, 143, $errno, $errstr, 30);
if(!$connection) return false;
$output = fgets($connection, 128); // banner
fputs($connection, "1 login $user $pass\r\n");
$output = fgets($connection, 128);
fputs($connection, "2 logout\r\n");
fclose($connection);
if (substr($output, 0, 4) == '1 OK') return true;
return false;
}
?>
lelkesa ¶
20 years ago
Note that - afaik - fgets reads a line until it reaches a line feed (\\n). Carriage returns (\\r) aren't processed as line endings.
However, nl2br insterts a <br /> tag before carriage returns as well.
This is useful (but not nice - I must admit) when you want to store a more lines in one.
<?php
function write_lines($text) {
$file = fopen('data.txt', 'a');
fwrite($file, str_replace("\n", ' ', $text)."\n");
fclose($file);
}
function read_all() {
$file = fopen('data.txt', 'r');
while (!feof($file)) {
$line = fgets($file);
echo '<u>Section</u><p>nl2br'.($line).'</p>';
}
fclose($file);
}
?>
Try it.
Anonymous ¶
20 years ago
If you need to simulate an un-buffered fgets so that stdin doesnt hang there waiting for some input (i.e. it reads only if there is data available) use this :
<?php
function fgets_u($pStdn) {
$pArr = array($pStdn);
if (false === ($num_changed_streams = stream_select($pArr, $write = NULL, $except = NULL, 0))) {
print("\$ 001 Socket Error : UNABLE TO WATCH STDIN.\n");
return FALSE;
} elseif ($num_changed_streams > 0) {
return trim(fgets($pStdn, 1024));
}
}
?>
svayn at yahoo dot com ¶
18 years ago
fgets is SLOW for scanning through large files. If you don't have PHP 5, use fscanf($file, "%s\n") instead.
hackajar <matt> yahoo <trot> com ¶
18 years ago
When working with VERY large files, php tends to fall over sideways and die.
Here is a neat way to pull chunks out of a file very fast and won't stop in mid line, but rater at end of last known line. It pulled a 30+ million line 900meg file through in ~ 24 seconds.
NOTE:
$buf just hold current chunk of data to work with. If you try "$buf .=" (note 'dot' in from of '=') to append $buff, script will come to grinding crawl around 100megs of data, so work with current data then move on!
//File to be opened
$file = "huge.file";
//Open file (DON'T USE a+ pointer will be wrong!)
$fp = fopen($file, 'r');
//Read 16meg chunks
$read = 16777216;
//\n Marker
$part = 0;
while(!feof($fp)) {
$rbuf = fread($fp, $read);
for($i=$read;$i > 0 || $n == chr(10);$i--) {
$n=substr($rbuf, $i, 1);
if($n == chr(10))break;
//If we are at the end of the file, just grab the rest and stop loop
elseif(feof($fp)) {
$i = $read;
$buf = substr($rbuf, 0, $i+1);
break;
}
}
//This is the buffer we want to do stuff with, maybe thow to a function?
$buf = substr($rbuf, 0, $i+1);
//Point marker back to last \n point
$part = ftell($fp)-($read-($i+1));
fseek($fp, $part);
}
fclose($fp);
dan at censornet dot com ¶
12 years ago
WARNING! fgets() and I presume any read() call to a file handle, e.g.
while(!feof(STDIN)) {
$line = fgets(STDIN);
...do something useful with $line...
}
...will result in a timeout after a default time of 60 seconds on my install. This behavior is non standard (not POSIX like) and seems to me to be a bug, or if not a major caveat which should be documented more clearly.
After the timeout fgets() will return FALSE (=== FALSE), however, you can check to see if the stream (file handle) has really closed by checking feof($stream), e.g.
while(!feof(STDIN)) {
$line = fgets(STDIN);
if($line === FALSE) {
if(feof(STDIN)) {
break;
}
continue;
}
...do something useful with $line...
}
bobo ¶
7 years ago
One easy sample, how to use "fgets":
This sample shows how to read the Information from an .txt file:
<?php
$z=fopen("protocol.txt","r"); //r = read
if($pointer!=false) //Is it possible to open the File?!
{
//Repeats all Information from protocol.txt
while(!feof($pointer)) //the loop runs till the Pointer is at the End of the File
{
$row = fgets($pointer); // $row reads the Information from the row of the File
echo "<p>".$row."</p>";
}
fclose($pointer); //File must be closed
}
else
{
echo"<p> It was not possible to open the File!</p>";
}
?>
bogdan at insightmed dot eu ¶
8 years ago
Even if this is not really related to PHP and its internals take care when using fgets for reading input from CLI on Linux systems as it may behave unexpected because of the limitations of arguments length on these systems. For example doing rtrim(fgets(STDIN), "\n") on a user input larger than 4095 characters will cut the the input string to 4095 characters. This shortcoming can be solved using "stty -icanon" before the script run, followed by a "stty icanon" after the script is run.
zsjpxah ¶
5 years ago
It's strange no one mentions "0" in this context.
Since "0" is considered to be false, a line with a single "0" can be treated as EOF if using the while assign idiom.
while ($line = fgets(STDIN, 2)) {
}
This may surprisingly break if a line starts with ")"
alex at alex-at dot ru ¶
9 years ago
For anyone who wants a proper non-blocking fgets for sockets, there is a tiny snippet that does just that (performance should be horrible compared to fgets though):
<?php
function read_line_nb($handle)
{
static $buffer = '';
static $lastOffset = 0;
$buffer .= fread($handle, 0x1000);
if (preg_match('#\\R#', $buffer, $m, PREG_OFFSET_CAPTURE, $lastOffset)) {
$line = substr($buffer, 0, $m[0][1] + strlen($m[0][0]));
$buffer = substr($buffer, $m[0][1] + strlen($m[0][0]));
return $line;
}
$lastOffset = strlen($buffer);
return false;
}
?>
wojons ¶
10 years ago
This goes out to Leigh Purdie (5 years ago) and also Dade Brandon (4 months ago)
So i say Leigh posting and though omg i need to change all my fgets to stream_get_line. Then i ran the tests as shown in Leigh Purdie comment His results:
$ time yes "This is a test line" | head -1000000 | php -r '$fp=fopen("php://stdin","r"); while($line=stream_get_line($fp,65535,"\n")) { 1; } fclose($fp);'
real 0m1.482s
user 0m1.616s
sys 0m0.152s
$ time yes "This is a test line" | head -1000000 | php -r '$fp=fopen("php://stdin","r"); while($line=fgets($fp,65535)) { 1; } fclose($fp);'
real 0m7.281s
user 0m7.392s
sys 0m0.136s
My Results:
$ time yes "This is a test line" | head -1000000 | php -r '$fp=fopen("php://stdin","r"); while($line=stream_get_line($fp,65535,"\n")) { 1; } fclose($fp);'
real 0m0.341s
user 0m0.352s
sys 0m0.148s
$ time yes "This is a test line" | head -1000000 | php -r '$fp=fopen("php://stdin","r"); while($line=fgets($fp,65535)) { 1; } fclose($fp);'
real 0m4.283s
user 0m4.128s
sys 0m0.448s
My results do show the same issue his results show. But first off PHP has at least gotten about 2-5 times faster then when the tests were first run (or better hardware).
Now to relate to Dade Brandon who states if you use a correct buffer size the perfomance is neck and neck.
$ time yes "This is a test line" | head -1000000 | php -r '$fp=fopen("php://stdin","r"); while($line=stream_get_line($fp,21,"\n")) { 1; } fclose($fp);'
real 0m0.336s
user 0m0.412s
sys 0m0.076s
$ time yes "This is a test line" | head -1000000 | php -r '$fp=fopen("php://stdin","r"); while($line=fgets($fp,21)) { 1; } fclose($fp);'
real 0m0.312s
user 0m0.364s
sys 0m0.192s
As you can see very close and fgets just coming just a little bit ahead. I suspect that fgets is reading backwards on the buffer or loads everything into its self then trys to figure it out where as a correct set buffer does the trick. Dade Brandon states that fgets lets you know how the line was delimited. stream_get_line lets you choose what you wanna call the delimiter using its 3rd option.
fgets has one more option that is important, you dont have to set the length of the line. So in a case where you may not know the length of the line maybe in handling Http protocol or something else like log lines you can simply leave it off and still get great performance.
$ time yes "This is a test line" | head -1000000 | php -r '$fp=fopen("php://stdin","r"); while($line=fgets($fp)) { 1; } fclose($fp);'
real 0m0.261s
user 0m0.260s
sys 0m0.232s
This is better then with a buffer set.
apardo at NOSPAM dot gmail dot com ¶
12 years ago
This is a a simple function to detect end of line type for any file.
<?php
function detectEndOfLine($file)
{
$handle = @fopen($file, "r");
if ($handle)
{
$char=0;
while (!$eol || feof($handle))
{
$char++;
$line = fgets($handle, $char);
$eol = preg_match("/(\r)+/", $line)? "W" : "";
if(!$eol) $eol = preg_match("/(\n)+/", $line)? "L" : "";
}
return $eol;
fclose($handle);
}
}
?>
david_sitller at blackbit dot de ¶
17 years ago
If you use the example from the command-description, i recommend to trim the $buffer for further use. The line feed ist still at the end of the line. I saw this when using PHP CLI.
Like this, checking a file-list for existing entries:
$handle = fopen ("/tmp/files.txt", "r");
while (!feof($handle)) {
$buffer = fgets($handle, 4096);
if (file_exists(rtrim($filename,"\n"))) {
echo $buffer;
} else {
echo $buffer." has been removed."
}
fclose ($handle);
hgs at cs dot columbia dot edu ¶
13 years ago
There seems to be an interaction between sockets and the auto_detect_line_endings setting that can cause rather peculiar behavior. Apparently, if the first line read from a socket is split across two TCP packets, the detector will look at the first TCP packet and determine that the system uses MacOS (\r) line endings, even though the LF is contained in the next packet. For example, this affected the PEAR Net_SMTP package, which would fail mysteriously for only some email servers.
jerem-NoSpam-ified at live dot com ¶
16 years ago
It's worth noting that this function only assumes chr(10) as a line break, but not chr(13). Personally, I prefer using chr(13) as a line break.
Anonymous ¶
10 years ago
This goes out to Leigh Purdie (5 years ago) and also Dade Brandon (4 months ago)
So i say Leigh posting and though omg i need to change all my fgets to stream_get_line. Then i ran the tests as shown in Leigh Purdie comment His results:
$ time yes "This is a test line" | head -1000000 | php -r '$fp=fopen("php://stdin","r"); while($line=stream_get_line($fp,65535,"\n")) { 1; } fclose($fp);'
real 0m1.482s
user 0m1.616s
sys 0m0.152s
$ time yes "This is a test line" | head -1000000 | php -r '$fp=fopen("php://stdin","r"); while($line=fgets($fp,65535)) { 1; } fclose($fp);'
real 0m7.281s
user 0m7.392s
sys 0m0.136s
My Results:
$ time yes "This is a test line" | head -1000000 | php -r '$fp=fopen("php://stdin","r"); while($line=stream_get_line($fp,65535,"\n")) { 1; } fclose($fp);'
real 0m0.341s
user 0m0.352s
sys 0m0.148s
$ time yes "This is a test line" | head -1000000 | php -r '$fp=fopen("php://stdin","r"); while($line=fgets($fp,65535)) { 1; } fclose($fp);'
real 0m4.283s
user 0m4.128s
sys 0m0.448s
My results do show the same issue his results show. But first off PHP has at least gotten about 2-5 times faster then when the tests were first run (or better hardware).
Now to relate to Dade Brandon who states if you use a correct buffer size the perfomance is neck and neck.
$ time yes "This is a test line" | head -1000000 | php -r '$fp=fopen("php://stdin","r"); while($line=stream_get_line($fp,21,"\n")) { 1; } fclose($fp);'
real 0m0.336s
user 0m0.412s
sys 0m0.076s
$ time yes "This is a test line" | head -1000000 | php -r '$fp=fopen("php://stdin","r"); while($line=fgets($fp,21)) { 1; } fclose($fp);'
real 0m0.312s
user 0m0.364s
sys 0m0.192s
As you can see very close and fgets just coming just a little bit ahead. I suspect that fgets is reading backwards on the buffer or loads everything into its self then trys to figure it out where as a correct set buffer does the trick. Dade Brandon states that fgets lets you know how the line was delimited. stream_get_line lets you choose what you wanna call the delimiter using its 3rd option.
fgets has one more option that is important, you dont have to set the length of the line. So in a case where you may not know the length of the line maybe in handling Http protocol or something else like log lines you can simply leave it off and still get great performance.
$ time yes "This is a test line" | head -1000000 | php -r '$fp=fopen("php://stdin","r"); while($line=fgets($fp)) { 1; } fclose($fp);'
real 0m0.261s
user 0m0.260s
sys 0m0.232s
This is better then with a buffer set.
