Parsing a text file into an array

[SEO Voter]

Just wanted to check what was the best and fastestway to upload a file and have the php script read each line of a text file into an array.

I will post my code after a few samples have been posted so as to not taint answers...

[Join us @ facebook]

The fastest way to parse a file is using something like :

Code:

<?php
$handle = fopen("path/to/file", "rb");
$contents = '';
while (!feof($handle)) {
  $contents .= fread($handle, 8192);
}
fclose($handle);
?>

The easiest is

Code:

<?php
$contents = file_get_contents("/path/to/file");
?>

I suppose the decision on what to use is depending on how big the file you use is(for memory reasons)

[Join us @ facebook]

I wasn't right. file_get_contents is faster than fread

from php.net :

Note:  If you just want to get the contents of a file into a string, use file_get_contents() as it has much better performance than the code above.

[SEO Voter]

Well, you have hit the nail on the head! It is exactly the same as what I have.

The part that parses the text file into an array (and in this case each line per element of the array) goes like this...

Code:

$array = explode("\n", $contents);

Nice!

[Join us @ facebook]

In this case I think the faster would be this :

Code:

<?php
$array = array();
$handle = fopen("path/to/file", "r");
while (!feof($handle)) {
  $array = fgets($handle, 4096);
}
fclose($handle);
?>

[SEO Voter]

fgets returns a string though, will this work?

I think that I will test both and time them!

[Join us @ facebook]

fgets($handle, 4096);

mean that will read one line from the $handle with maximum limit 4096 bytes/line.

It is much faster, because it will create the array while parsing, instead of explode("\n", file_get_contents($file)); which will first load the file in memory and then will parse it to create the array.

[Join us @ facebook]

Ooops I made a small error in the code.

The right is $array [] = fgets($handle, 4096);

[SEO Voter]

Yes, I suspect you are right! I will time both and post results. It will be a nice test to show...

[Join us @ facebook]

Yeah. I wait for your results

[SEO Voter]

You may be pleasantly surprised here! Here are the results run five times...

explode: 0.043 ms
fgets:     0.131 ms

explode: 0.043 ms
fgets:     0.132 ms

explode: 0.041 ms
fgets:     0.134 ms

explode: 0.043 ms
fgets:     0.133 ms

explode: 0.040 ms
fgets:     0.132 ms

file size : 1.6Mb

...

[SEO Voter]

On average it seems to be three times faster... in explodes favour!

[Join us @ facebook]

Strange huh?

You tried it in windows or linux?

[SEO Voter]

at this stage it is still on my windows laptop, but I will upload it to our linux server and test again...

[Join us @ facebook]

You should do that.

Also I think  there is another catch in this. Even that explode is faster, it costs more in memory.

On a server environment where many clients can seek the same page at once maybe this could be a prob.

Of course all of these are just predictions, and are actually fitable for sites with hundrends of thousands daily visits.

[SEO Voter]

I got very similar results on a linux server.

I was also thinking about the load on the server, but i think that it is too much effort to monitor that.

The files that are going to be uploaded shouldn't be more than 100K anyway, so either method will do fine anyway!

Anyway, well at least we learnt something today!

[Join us @ facebook]

Anyway, well at least we learnt something today!

That's true, and as my ancient fathers said once "Girasko aei didaskomenos" which mean something like "I learn while I am getting older(forever)"

[Guest]

[36.302]

Googlebot, Yahoo crawler, Baidu Spider