file_put_contents
on file() and flock()
My supervisor came up with a brilliant plan to workaround the inability of the file() to work on a flock()'ed file.
We created a dummy file called lockfile.txt. We would flock() lockfile.txt. Once we had a lock on it, we used file() on the file we wanted to read, then altered the file and called fclose on both files.
file
(PHP 4, PHP 5)
file — 把整个文件读入一个数组中
说明
array file
( string $filename
[, int $use_include_path
[, resource $context
]] )
和 file_get_contents() 一样，只除了 file() 将文件作为一个数组返回。数组中的每个单元都是文件中相应的一行，包括换行符在内。如果失败 file() 返回 FALSE。
如果也想在 include_path 中搜寻文件的话，可以将可选参数 use_include_path 设为 "1"。
<?php
// 将一个文件读入数组。本例中通过 HTTP 从 URL 中取得 HTML 源文件。
$lines = file('http://www.example.com/');
// 在数组中循环，显示 HTML 的源文件并加上行号。
foreach ($lines as $line_num => $line) {
echo "Line #<b>{$line_num}</b> : " . htmlspecialchars($line) . "<br />\n";
}
// 另一个例子将 web 页面读入字符串。参见 file_get_contents()。
$html = implode('', file ('http://www.example.com/'));
?>
Tip
如果“fopen wrappers”已经被激活，则在本函数中可以把 URL 作为文件名来使用。请参阅 fopen() 函数来获取怎样指定文件名的详细信息以及支持 URL 封装协议的列表：支持的协议／封装协议列表。
Note: 返回的数组中每一行都包括了行结束符，因此如果不需要行结束符时还需要使用 rtrim() 函数。
Note: 如果碰到 PHP 在读取文件时不能识别 Macintosh 文件的行结束符，可以激活 auto_detect_line_endings 运行时配置选项。
Note: 从 PHP 4.3.0 开始可以用 file_get_contents() 来将文件读入到一个字符串返回。
从 PHP 4.3.0 开始 file() 可以安全用于二进制文件。
Note: 对 context 的支持是 PHP 5.0.0 添加的。有关 context 的说明见Streams。
Warning
当使用 SSL 时，Microsoft IIS 将违反协议不发送 close_notify 标记就关闭连接。PHP 将在到达数据尾端时报告 "SSL: Fatal Protocol Error"。要绕过此问题，应将 error_reporting 级别降低为不包括警告。PHP 4.3.7 及更高版本可以在当使用 https:// 封装协议打开流的时候检测出有此问题的 IIS 服务器并抑制警告。如果使用 fsockopen() 来创建一个 ssl:// 套接字，则需要自己检测并抑制警告信息。
参见 readfile()，fopen()，fsockopen()，popen()，file_get_contents() 和 include()。
add a note
User Contributed Notesfile
PenguinMan98 at usa dot net
21-Apr-2008 10:49
21-Apr-2008 10:49
jon+spamcheck at phpsitesolutions dot com
16-Apr-2008 08:03
16-Apr-2008 08:03
A user suggested using rtrim always, due to the line ending conflict with files that have an EOL that differs from the server EOL.
Using rtrim with it's default character replacement is a bad solution though, as it removes all whitespace in addition to the '\r' and '\n' characters.
A good solution using rtrim follows:
<?php
$line = rtrim($line, "\r\n") . PHP_EOL;
?>
This removes only EOL characters, and replaces with the server's EOL character, thus making preg_* work fine when matching the EOL ($)
bubbles
05-Apr-2008 10:45
05-Apr-2008 10:45
althought it's mentioned twice in the description, it took me a whole night to figure out why i got new-lines in my array.
hence you have to put a flag on it FILE_IGNORE_NEW_LINES.
I mention, just you wouldn't miss this little anoying thing.
*you can use trim, but it's slighty different.
vbchris at gmail dot com
16-Feb-2008 09:15
16-Feb-2008 09:15
If you're getting "failed to open stream: Permission denied" when trying to use either file() or fopen() to access files on another server. Check your host doesn't have any firewall restrictions in-place which prevent outbound connections. This is the case with my host Aplus.net
Reversed: moc dot liamg at senroc dot werdna
12-Jul-2007 09:25
12-Jul-2007 09:25
This note applies to PHP 5.1.6 under Windows (although may apply to other versions).
It appears that the 'FILE_IGNORE_NEW_LINES' flag doesn't remove newlines properly when reading Windows-style text files, i.e. files whose lines end in '\r\n'.
Solution: Always use 'rtrim()' in preference to 'FILE_IGNORE_NEW_LINES'.
info at carstanje dot com
28-Nov-2006 08:33
28-Nov-2006 08:33
Using file() for reading large text files > 10 Mb gives problems, therefore you should use this instead. It is much slower but it works fine. $lines will return an array with all the lines.
<?php
$handle = @fopen('yourfile...', "r");
if ($handle) {
while (!feof($handle)) {
$lines[] = fgets($handle, 4096);
}
fclose($handle);
}
?>
richardtcunningham at gmail dot com
11-Jul-2006 09:19
11-Jul-2006 09:19
justin at visunet dot ie's note of 20-Mar-2003 states
"Note: Now that file() is binary safe it is 'much' slower than it used to be. If you are planning to read large files it may be worth your while using fgets() instead of file()."
I tested fgets(), file_get_contents(), and file() on PHP 4.3.2 and PHP 5 and timed each to be under a second with over 200,000 lines. I do not know if he was testing extremely long lines or what, but I could not duplicate the difference that he mentioned.
jonathan dot gotti at free dot fr
01-Feb-2006 10:52
01-Feb-2006 10:52
you can use
$file = array_map('rtrim',file('myfile.txt'));
to remove annoying ending lines of the resulting array.
Nuts
18-Jan-2006 11:16
18-Jan-2006 11:16
WARNING ON WINDOWS:
file() function will add "\r\n" in to the end of the row, even if you use only "\n" char to make rows in the file!
On UNIX systems there is no such problem.
dir @ badblue com
12-Sep-2003 09:48
12-Sep-2003 09:48
Jeff's array2file function is a good start; here are a couple of improvements (no possibility of handle leak when fwrite fails, additional capability of both string2file and array2file; presumably faster performance through use of implode).
<?php
function String2File($sIn, $sFileOut) {
$rc = false;
do {
if (!($f = fopen($sFileOut, "wa+"))) {
$rc = 1; break;
}
if (!fwrite($f, $sIn)) {
$rc = 2; break;
}
$rc = true;
} while (0);
if ($f) {
fclose($f);
}
return ($rc);
}
function Array2File($aIn, $sFileOut) {
return (String2File(implode("\n", $aIn), $sFileOut));
}
?>
If you're generating your string text using a GET or POST from a TEXTAREA (e.g., a mini-web-text-editor), remember that strip_slashes and str_replace of "/r/n" to "/n" may be necessary as well using these functions.
HTH --dir @ badblue com
John
20-Jul-2003 11:32
20-Jul-2003 11:32
after many months of confusion and frustration, i have finally figured out something that i should have noticed the first time around.
you can't file("test.txt") when that same file has been flocked. i guess i didn't have a full understanding of what i was doing when i used flock(). all i had to do was move the flock() around, and all was well.
justin at visunet dot ie
20-Mar-2003 05:36
20-Mar-2003 05:36
Note: Now that file() is binary safe it is 'much' slower than it used to be. If you are planning to read large files it may be worth your while using fgets() instead of file() For example:
<?php
$fd = fopen ("log_file.txt", "r");
while (!feof ($fd))
{
$buffer = fgets($fd, 4096);
$lines[] = $buffer;
}
fclose ($fd);
?>
The resulting array is $lines.
I did a test on a 200,000 line file. It took seconds with fgets() compared to minutes with file().
andrea at brancatelli dot it
16-Mar-2002 07:16
16-Mar-2002 07:16
file() has a strange behaviour when reading file with both \n and \r as line delimitator (DOS files), since it will return an array with every single line but with just a \n in the end. It seems like \r just disappears.
This is happening with PHP 4.0.4 for OS/2. Don't know about the Windows version.
php@don't_spam_me
09-Feb-2002 08:56
09-Feb-2002 08:56
It appears that the file() function causes file access problems for perl cgi scripts accessing the same files. I am using Perl v5.6.0 in linux with PHP/4.0.4pl1. After running a php app using the file() function, any perl cgi trying to access the same file randomly dies returning an internal server error: premature end of script headers.
The simple fix is to use fopen(), fgets() and fclose() instead of file().