Fastest way to reduce a csv file to two columns - php

Question

I have a csv file that has like 30,000 rows in it. It also has like 9 columns. In the interest of speeding up the processing of everything I want to reduce the file to the two columns that I need and remove the rest. here is what I have done.
$retardment=1;//17;// 151; //499;// 991;// 1877
if (($handle = fopen($source, "r")) !== FALSE) {
$stock_handle = fopen($source_stock, "w+");
$row=0;
$col=array();
while (($line = fgetcsv($handle, 100000, ",")) !== FALSE) {
unset($line[1]);
unset($line[2]);
unset($line[3]);
unset($line[4]);
unset($line[5]);
unset($line[6]);
unset($line[8]);
unset($line[9]);
if($row%$retardment<1){
fputcsv($stock_handle, $line);
}
unset($line);
$row++;
}
fclose($handle);
fclose($stock_handle);
}
I am coping it to a new file and this works... but it seems to be pretty slow. Any ideas on how to make it faster? Thank you for the help.
Cheers -Jeremy
{EDIT}
So far this seems to take just as long. But works just fine
while (($line = fgetcsv($handle, 100000, ",")) !== FALSE) {
if($row%$retardment<1){
fputcsv($stock_handle, array($line[0],$line[7]));
}
$row++;
}

Answer 1

You could replace those unset() calls with...
$line = array($line[0], $line[7]);
Alternatively, remember that unset() takes multiple arguments...
unset($line[1], $line[2], ...);

Answer 2

You can speed it up fractionally more, but again it's a microtime()-measurable improvement: perception is that it won't be noticeably faster.
while (($line = fgetcsv($handle, 100000, ",")) !== FALSE) {
if($row++ % $retardment < 1){
fputcsv($stock_handle, array($line[0],$line[7]));
}
}
but as your script is IO-bound, it's the actual reads and writes that are the slowest functions, and you can't speed those up.
Using stream_copy_to_stream() with a stream input filter might be another approach, but you won't see much noticeable improvement unless you can reduce disk access times