I'm inserting chunks of data on a regular basis, like this:

$query = "INSERT INTO table (pubdate, title, link, description) VALUES ";

foreach ($array as $key) {
 list($var_1, $var_2, $var_3, $var_4) = explode('|:|', $key);

 $query .= sprintf("('%s', '%s', '%s', '%s', '%s'), ",
  mysql_real_escape_string($var_1),
  mysql_real_escape_string($var_2),
  mysql_real_escape_string($var_3),
  mysql_real_escape_string($var_4));
}

$query = rtrim(', ', $query);

mysql_query($query);

(that code is still in development phase, so please ignore any typos)

But now, I want to modify it so that it updates duplicate keys, which I assume uses something like:

$query .= "ON DUPLICATE KEY UPDATE title='%s', link='%s', description='%s'"

But since I'm using literal replacements, and I'm doing the VALUES in a loop, I'm not sure where to put this.

The only option I can think of is to create a long series of queries within the loop:

$query = '';

foreach ($array as $key) {
 list($var_1, $var_2, $var_3, $var_4) = explode('|:|', $key);

 $query .= "INSERT INTO table (pubdate, title, link, description) VALUES ";
 $query .= sprintf("('%s', '%s', '%s', '%s', '%s')",
  mysql_real_escape_string($var_1),
  mysql_real_escape_string($var_2),
  mysql_real_escape_string($var_3),
  mysql_real_escape_string($var_4));

 $query .= "ON DUPLICATE KEY UPDATE title='%s', link='%s', description='%s';\n"
}

mysql_query($query);

I could be inserting 100 lines at once, though. Surely that's not the best solution?