php remove duplicate if same mobile phone number is in .csv file - php

Question

I have a separate php script file that saves to file a number csv values via a html contact form.
I would like a maximum of 2 duplicate rows based on mobile phone entries in csv file,
any more and I would want the current record deleted.
I am using the $_GET()(no $_POST() functions) function to record all entries, and then save to file.
Im just having issues with deleting duplicates if the mobile number is already TWICE in the file.
Any help would be greatly appreciated.
**ADDED MORE CODE AND COMMENT BELOW**
I have edited the code, but I am still running into trouble with removing duplicates, let alone check for 2 dupes first. I will do the sanitize and better code 'after' I have some function (help!).
Thanks again for your help :)
<?php
$filename = "input.csv";
$csv_output .= "\n";$title=$_GET[title];$fname=$_GET[fname];
$sname=$_GET[sname];$notes=$_GET[notes];$mobile=$_GET[mobile];
$string="$title,$fname,$sname,$mobile,$notes,$csv_output";
$file = fopen($filename, "c");
// see details on the 'c' mode here http://us3.php.net/manual/en/function.fopen.php - it will create a file if it does not exist.
// Now acquire an exclusive via flock() - http://us3.php.net/manual/en/function.flock.php
flock($file, LOCK_EX); // this will block till some other reader/writer has released the lock.
$stat = fstat($file)
if($stat['size'] == 0) {
// file created for the first time
fwrite($file, "Title,First Name,Last Name,MobileNumber,Notes\n$string");
flock($file, LOCK_UN);
fclose($file);
return;
}
// File not empty - scan thru line by line via fgets(), and detect duplicates
// If a duplicate is detected, just flock($file, LOCK_UN), close the file and return - ///// no need to fwrite the line.
while (($buffer = fgets($file, 2188)) !== false) {
if(!stripos($buffer, ",$mobile,") {
$mobile .= $buffer;
}
else {
flock($file, LOCK_UN);
fclose($file);
return;
}
}
?>

Answer 1

Are you running this on a Linux/Unix system? If so, the way you have accessed the file will lead to race-conditions and possible corruption of the file.
You need to ensure that the write to the file is done in a serialized manner if multiple processes are attempting to write to the same file.
As you don't want to explore other alternatives like a db (even key-value file-based dbs), a pseudo-code approach is:
$file = fopen($filename, "c"); // see details on the 'c' mode here http://us3.php.net/manual/en/function.fopen.php - it will create a file if it does not exist.
// Now acquire a exclusive via flock() - http://us3.php.net/manual/en/function.flock.php
flock($file, LOCK_EX); // this will block till some other reader/writer has released the lock.
$stat = fstat($file)
if($stat['size'] == 0)
{
// file created for the first time
fwrite($file, "Title,First Name,Last Name,MobileNumber,Notes\n$string");
flock($file, LOCK_UN);
fclose($file);
return;
}
// File not empty - scan thru line by line via fgets(), and detect duplicates
// If a duplicate is detected, just flock($file, LOCK_UN), close the file and return - no need to fwrite the line.
// Otherwise fwrite() the line
.
.
flock($file, LOCK_UN);
fclose($file);
You can fill in the details in the middle part - hope you got the gist of it.
You could potentially make it more 'scalable' by initially grabbing a read lock (this will allow multiple readers to run concurrently, and only the writer will block). Once the read portion is done, you need to release the lock, and if a write needs to be done (i.e. no duplicates detected), then grab a writer lock etc...
Clearly this is not a ideal solution but if your file contents are going to be small, it may suffice.
Stating the obvious, you would need to do better error handling with all file-based operations.
A tangential point: you should also sanitize the data from $_GET before going to the core logic to catch for invalid inputs.
Hope this helps.