How do I read a file backwards line by line using fseek?
code can be helpful. must be cross platform and pure php.
many thanks in advance
regards
Jera
How do I read a file backwards line by line using fseek?
code can be helpful. must be cross platform and pure php.
many thanks in advance
regards
Jera
The question is asking using fseek, so can only assume that performance is an issue and file() is not the solution. Here is a simple approach using fseek:
My file.txt
#file.txt
Line 1
Line 2
Line 3
Line 4
Line 5
And the code:
<?php
$fp = fopen('file.txt', 'r');
$pos = -2; // Skip final new line character (Set to -1 if not present)
$lines = array();
$currentLine = '';
while (-1 !== fseek($fp, $pos, SEEK_END)) {
$char = fgetc($fp);
if (PHP_EOL == $char) {
$lines[] = $currentLine;
$currentLine = '';
} else {
$currentLine = $char . $currentLine;
}
$pos--;
}
$lines[] = $currentLine; // Grab final line
var_dump($lines);
Output:
array(5) {
[0]=>
string(6) "Line 5"
[1]=>
string(6) "Line 4"
[2]=>
string(6) "Line 3"
[3]=>
string(6) "Line 2"
[4]=>
string(6) "Line 1"
}
You don't have to append to the $lines array like I am, you can print the output straight away if that is the purpose of your script. Also it is easy to introduce a counter if you want to limit the number of lines.
$linesToShow = 3;
$counter = 0;
while ($counter <= $linesToShow && -1 !== fseek($fp, $pos, SEEK_END)) {
// Rest of code from example. After $lines[] = $currentLine; add:
$counter++;
}
If you are going to read the entire file in anyways, just use file() to read the file in as an array (each line is each element in the array) and then use array_reverse() to flip the array backwards and loop through that. Or just do a reverse for loop where you start at the end and decrement on each loop.
$file = file("test.txt");
$file = array_reverse($file);
foreach($file as $f){
echo $f."<br />";
}
$i >= 0
though. –
Numbles array_slice
is really only related to array_reverse
in that they both work on arrays. They don't really do the same thing at all. The array_reverse
here was just for readability. If you didn't want it, since file
produces a numeric array you could just do a for loop decrementing your iterator like for($i=count($file)-1; $i>=0; $i--){ $line = $file[$i]; /*do something*/ }
. –
Numbles The question is asking using fseek, so can only assume that performance is an issue and file() is not the solution. Here is a simple approach using fseek:
My file.txt
#file.txt
Line 1
Line 2
Line 3
Line 4
Line 5
And the code:
<?php
$fp = fopen('file.txt', 'r');
$pos = -2; // Skip final new line character (Set to -1 if not present)
$lines = array();
$currentLine = '';
while (-1 !== fseek($fp, $pos, SEEK_END)) {
$char = fgetc($fp);
if (PHP_EOL == $char) {
$lines[] = $currentLine;
$currentLine = '';
} else {
$currentLine = $char . $currentLine;
}
$pos--;
}
$lines[] = $currentLine; // Grab final line
var_dump($lines);
Output:
array(5) {
[0]=>
string(6) "Line 5"
[1]=>
string(6) "Line 4"
[2]=>
string(6) "Line 3"
[3]=>
string(6) "Line 2"
[4]=>
string(6) "Line 1"
}
You don't have to append to the $lines array like I am, you can print the output straight away if that is the purpose of your script. Also it is easy to introduce a counter if you want to limit the number of lines.
$linesToShow = 3;
$counter = 0;
while ($counter <= $linesToShow && -1 !== fseek($fp, $pos, SEEK_END)) {
// Rest of code from example. After $lines[] = $currentLine; add:
$counter++;
}
<?php
class ReverseFile implements Iterator
{
const BUFFER_SIZE = 4096;
const SEPARATOR = "\n";
public function __construct($filename)
{
$this->_fh = fopen($filename, 'r');
$this->_filesize = filesize($filename);
$this->_pos = -1;
$this->_buffer = null;
$this->_key = -1;
$this->_value = null;
}
public function _read($size)
{
$this->_pos -= $size;
fseek($this->_fh, $this->_pos);
return fread($this->_fh, $size);
}
public function _readline()
{
$buffer =& $this->_buffer;
while (true) {
if ($this->_pos == 0) {
return array_pop($buffer);
}
if (count($buffer) > 1) {
return array_pop($buffer);
}
$buffer = explode(self::SEPARATOR, $this->_read(self::BUFFER_SIZE) . $buffer[0]);
}
}
public function next()
{
++$this->_key;
$this->_value = $this->_readline();
}
public function rewind()
{
if ($this->_filesize > 0) {
$this->_pos = $this->_filesize;
$this->_value = null;
$this->_key = -1;
$this->_buffer = explode(self::SEPARATOR, $this->_read($this->_filesize % self::BUFFER_SIZE ?: self::BUFFER_SIZE));
$this->next();
}
}
public function key() { return $this->_key; }
public function current() { return $this->_value; }
public function valid() { return ! is_null($this->_value); }
}
$f = new ReverseFile(__FILE__);
foreach ($f as $line) echo $line, "\n";
To completely reverse a file:
$fl = fopen("\some_file.txt", "r"); for($x_pos = 0, $output = ''; fseek($fl, $x_pos, SEEK_END) !== -1; $x_pos--) { $output .= fgetc($fl); } fclose($fl); print_r($output);
Of course, you wanted line-by-line reversal...
$fl = fopen("\some_file.txt", "r"); for($x_pos = 0, $ln = 0, $output = array(); fseek($fl, $x_pos, SEEK_END) !== -1; $x_pos--) { $char = fgetc($fl); if ($char === "\n") { // analyse completed line $output[$ln] if need be $ln++; continue; } $output[$ln] = $char . ((array_key_exists($ln, $output)) ? $output[$ln] : ''); } fclose($fl); print_r($output);
Really though, Jonathan Kuhn has the best answer IMHO above. The only cases you'd not use his answer that I know of is if file
or like functions are disabled via php.ini, yet the admin forgot about fseek, or when opening a huge file just get the last few lines of contents would magically save memory this way.
Note: Error handling not included. And, PHP_EOL didn't cooperate, so I used "\n" to denote end of line instead. So, above may not work in all cases.
You cannot fseek line by line, because you do not know how long the lines are until you read them.
You should either read the whole file into a list of lines, or if the file is too big for that and you only need the last lines, read fixed-sized chunks from the end of the file and implement a bit more complicated logic which detects lines from such data.
Reading the entire file into an array and reversing is fine unless the file is enormous.
You could perform a buffered read of your file from back to front with something like this:
This code read file backwards. This code ignore modifications on reading, example apache access.log new lines on procressing.
$f = fopen('FILE', 'r');
fseek($f, 0, SEEK_END);
$pos = ftell($f);
$pos--;
while ($pos > 0) {
$chr = fgetc($f);
$pos --;
fseek($f, $pos);
if ($chr == PHP_EOL) {
YOUR_OWN_FUNCTION($rivi);
$rivi = NULL;
continue;
}
$rivi = $chr.$rivi;
}
fclose($f);
Here's a drop in replacement(ish) for fgets($fp) called fgetsr() that reads lines from a file in reverse order.
This code is verbatim so you should (famous last words) be able to copy it into a file on your server and run it. Though you may well need to change the filename in the fopn() call.
<?php
header('Content-Type: text/plain');
$fp = fopen('post.html', 'r');
while($line = fgetsr($fp)) {
echo $line;
}
// Read a line from the file but starting from the end
//
// @param $fp integer The file pointer
//
function fgetsr($fp)
{
// Make this variable persistent inside this function
static $seeked;
// The line buffer that will eventually be returned
$line = '';
// Initially seek to the end of the file
if (!$seeked) {
fseek($fp, -1, SEEK_END);
$seeked = true;
}
// Loop through all of the characters in the file
while(strlen($char = fgetc($fp)) {
// fgetc() advances that pointer so go back TWO places
// instead of one
fseek($fp, -2, SEEK_CUR);
//
// Check for a newline (LF). If a newline is found
// then break out of the function and return the
// line that's stored in the buffer.
//
// NB The first line in the file (ie the last to
// be read)has a special case
//
if (ftell($fp) <= 0) {
fseek($fp, 0, SEEK_SET);
$line = fgets($fp);
fseek($fp, 0, SEEK_SET);
return $line;
} else if ($char === "\n") {
$line = strrev($line);
return $line . "\n";
} else {
$line .= $char;
}
}
}
?>
Functions to read a file line-by-line in reverse:
function revfopen($filepath, $mode)
{
$fp = fopen($filepath, $mode);
fseek($fp, -1, SEEK_END);
if (fgetc($fp) !== PHP_EOL) {
fseek($fp, 1, SEEK_END);
}
return $fp;
}
function revfgets($fp)
{
$s = '';
while (true) {
if (fseek($fp, -2, SEEK_CUR) === -1) {
return false;
}
if (($c = fgetc($fp)) === PHP_EOL) {
break;
}
$s = $c . $s;
}
return $s;
}
Example use case: parse a long file until some date:
$fp = revfopen('/path/to/file', 'r');
$buffer = '';
while (($line = revfgets($fp)) !== false) {
if (strpos($line, '05-10-2021') === 0) {
break;
}
array_unshift($buffer, $line);
}
echo implode("\n", $buffer);
I know this has been answered already but I found another, maybe faster, way.
// Read last 5000 chars of 'foo.log'
if(file_exists('foo.log') && $file = fopen('foo.log', 'r')) {
fseek($file, -5000, SEEK_END);
$text = stream_get_line($file, 5000);
var_dump($text);
fclose($file);
}
© 2022 - 2024 — McMap. All rights reserved.