Export to CSV via PHP

I personally use this function to create CSV content from any array.

function array2csv(array &$array)
{
   if (count($array) == 0) {
     return null;
   }
   ob_start();
   $df = fopen("php://output", 'w');
   fputcsv($df, array_keys(reset($array)));
   foreach ($array as $row) {
      fputcsv($df, $row);
   }
   fclose($df);
   return ob_get_clean();
}

Then you can make your user download that file using something like:

function download_send_headers($filename) {
    // disable caching
    $now = gmdate("D, d M Y H:i:s");
    header("Expires: Tue, 03 Jul 2001 06:00:00 GMT");
    header("Cache-Control: max-age=0, no-cache, must-revalidate, proxy-revalidate");
    header("Last-Modified: {$now} GMT");

    // force download  
    header("Content-Type: application/force-download");
    header("Content-Type: application/octet-stream");
    header("Content-Type: application/download");

    // disposition / encoding on response body
    header("Content-Disposition: attachment;filename={$filename}");
    header("Content-Transfer-Encoding: binary");
}

Usage example:

download_send_headers("data_export_" . date("Y-m-d") . ".csv");
echo array2csv($array);
die();

You can export the date using this command.

<?php

$list = array (
    array('aaa', 'bbb', 'ccc', 'dddd'),
    array('123', '456', '789'),
    array('"aaa"', '"bbb"')
);

$fp = fopen('file.csv', 'w');

foreach ($list as $fields) {
    fputcsv($fp, $fields);
}

fclose($fp);
?>

First you must load the data from the mysql server in to a array

Just for the record, concatenation is waaaaaay faster (I mean it) than fputcsv or even implode; And the file size is smaller:

// The data from Eternal Oblivion is an object, always
$values = (array) fetchDataFromEternalOblivion($userId, $limit = 1000);

// ----- fputcsv (slow)
// The code of @Alain Tiemblo is the best implementation
ob_start();
$csv = fopen("php://output", 'w');
fputcsv($csv, array_keys(reset($values)));
foreach ($values as $row) {
    fputcsv($csv, $row);
}
fclose($csv);
return ob_get_clean();

// ----- implode (slow, but file size is smaller)
$csv = implode(",", array_keys(reset($values))) . PHP_EOL;
foreach ($values as $row) {
    $csv .= '"' . implode('","', $row) . '"' . PHP_EOL;
}
return $csv;
// ----- concatenation (fast, file size is smaller)
// We can use one implode for the headers =D
$csv = implode(",", array_keys(reset($values))) . PHP_EOL;
$i = 1;
// This is less flexible, but we have more control over the formatting
foreach ($values as $row) {
    $csv .= '"' . $row['id'] . '",';
    $csv .= '"' . $row['name'] . '",';
    $csv .= '"' . date('d-m-Y', strtotime($row['date'])) . '",';
    $csv .= '"' . ($row['pet_name'] ?: '-' ) . '",';
    $csv .= PHP_EOL;
}
return $csv;

This is the conclusion of the optimization of several reports, from ten to thousands rows. The three examples worked fine under 1000 rows, but fails when the data was bigger.