It would be easier if you just did a quick true/false check in the SQL and checked the flag that came back.
$sql = "SELECT "
. "(SELECT 1 FROM `users` WHERE `username` = '" . mysql_real_escape_string($username) . "'), "
. "(SELECT 1 FROM `users` WHERE `email` = '" . mysql_real_escape_string($email) . "')";
$query = mysql_query($sql);
if (mysql_num_rows($query) > 0) {
$foundFlags = mysql_fetch_assoc($query);
if ($foundFlags['username']) {
$error[] = "username is existing";
}
if ($foundFlags['email']) {
$error[] = "email is existing";
}
} else {
// General error as the query should always return
}
When it does not find an entry, it will return NULL in the flag, which evaluates to false, so the if
condition is fine.
Note that you could generalise it for a field list like this:
$fieldMatch = array('username' => $username, 'email' => $email);
$sqlParts = array();
foreach ($fieldMatch as $cFieldName => $cFieldValue) {
$sqlParts[] = "(SELECT 1 FROM `users` WHERE `" . $cFieldName . "` = '" . mysql_real_escape_string($cFieldValue) . "')";
}
$sql = "SELECT " . implode(", ", $sqlParts);
$query = mysql_query($sql);
if (mysql_num_rows($query) > 0) {
$foundFlags = mysql_fetch_assoc($query);
foreach ($foundFlags as $cFieldName => $cFlag) {
if ($foundFlags[$cFieldName]) {
$error[] = $cFieldName . " is existing";
}
}
} else {
// General error as the query should always return
}
NB. Note that assumes all fields are strings, or other string-escaped types (eg. date/time).