/**
* Stuff that only needs to be done once. Builds regular expressions and sorts the reserved words.
*/
protected static function init()
{
if (self::$init) {
return;
}
// Sort reserved word list from longest word to shortest
usort(self::$reserved, array('SqlFormatter', 'sortLength'));
// Set up regular expressions
self::$regex_boundaries = '(' . implode('|', array_map(array('SqlFormatter', 'quote_regex'), self::$boundaries)) . ')';
self::$regex_reserved = '(' . implode('|', array_map(array('SqlFormatter', 'quote_regex'), self::$reserved)) . ')';
self::$regex_reserved_toplevel = str_replace(' ', '\\s+', '(' . implode('|', array_map(array('SqlFormatter', 'quote_regex'), self::$reserved_toplevel)) . ')');
self::$regex_reserved_newline = str_replace(' ', '\\s+', '(' . implode('|', array_map(array('SqlFormatter', 'quote_regex'), self::$reserved_newline)) . ')');
self::$regex_function = '(' . implode('|', array_map(array('SqlFormatter', 'quote_regex'), self::$functions)) . ')';
self::$init = true;
}
/**
* Stuff that only needs to be done once. Builds regular expressions and sorts the reserved words.
*/
protected static function init()
{
if (self::$init) {
return;
}
// Sort reserved word list from longest word to shortest, 3x faster than usort
$reservedMap = array_combine(self::$reserved, array_map('strlen', self::$reserved));
arsort($reservedMap);
self::$reserved = array_keys($reservedMap);
// Set up regular expressions
self::$regex_boundaries = '(' . implode('|', array_map(array(__CLASS__, 'quote_regex'), self::$boundaries)) . ')';
self::$regex_reserved = '(' . implode('|', array_map(array(__CLASS__, 'quote_regex'), self::$reserved)) . ')';
self::$regex_reserved_toplevel = str_replace(' ', '\\s+', '(' . implode('|', array_map(array(__CLASS__, 'quote_regex'), self::$reserved_toplevel)) . ')');
self::$regex_reserved_newline = str_replace(' ', '\\s+', '(' . implode('|', array_map(array(__CLASS__, 'quote_regex'), self::$reserved_newline)) . ')');
self::$regex_function = '(' . implode('|', array_map(array(__CLASS__, 'quote_regex'), self::$functions)) . ')';
self::$init = true;
}
/**
* Return the next token and token type in a SQL string.
* Quoted strings, comments, reserved words, whitespace, and punctuation are all their own tokens.
*
* @param String $string The SQL string
* @param array $previous The result of the previous getNextToken() call
*
* @return Array An associative array containing a 'token' and 'type' key.
*/
protected static function getNextToken($string, $previous = null)
{
// If the next token is a comment
if ($string[0] === '#' || substr($string, 0, 2) === '--' || substr($string, 0, 2) === '/*') {
// Comment until end of line
if ($string[0] === '-' || $string[0] === '#') {
$last = strpos($string, "\n");
$type = 'comment';
} else {
// Comment until closing comment tag
$last = strpos($string, "*/", 2) + 2;
$type = 'block comment';
}
if ($last === false) {
$last = strlen($string);
}
return array('token' => substr($string, 0, $last), 'type' => $type);
}
// If the next item is a string
if (in_array($string[0], self::$quotes)) {
$quote = $string[0];
for ($i = 1, $length = strlen($string); $i < $length; $i++) {
$next_char = null;
if (isset($string[$i + 1])) {
$next_char = $string[$i + 1];
}
// Escaped (either backslash or backtick escaped)
if ($quote !== '`' && $string[$i] === '\\' || $quote === '`' && $string[$i] === '`' && $next_char === '`') {
$i++;
} elseif ($string[$i] === $quote) {
break;
}
}
if ($quote === '`') {
$type = 'backtick quote';
} else {
$type = 'quote';
}
return array('token' => substr($string, 0, $i + 1), 'type' => $type);
}
// Separators
if (in_array($string[0], self::$boundaries)) {
// If it is a simple string or empty between the parentheses, just count as a word
// this makes it so we don't split things like NOW() or COUNT(*) into separate lines
if ($string[0] === '(') {
// "()"
if (isset($string[1]) && $string[1] === ')') {
return array('token' => '()', 'type' => 'word');
}
// "(word/whitespace/boundary)"
$next_token = self::getNextToken(substr($string, 1));
$length = strlen($next_token['token']);
if (isset($string[$length + 1]) && $string[$length + 1] === ')') {
if ($next_token['type'] === 'word' || $next_token['type'] === 'whitespace' || $next_token['type'] === 'boundary') {
return array('token' => '(' . $next_token['token'] . ')', 'type' => 'word');
}
}
}
//return single parentheses as their own token
if ($string[0] === '(' || $string[0] === ')') {
return array('token' => $string[0], 'type' => $string[0]);
}
// If there are 1 or more boundary characters together, return as a single word
$next_token = self::getNextToken(substr($string, 1));
if ($next_token['type'] === 'boundary') {
return array('token' => $string[0] . $next_token['token'], 'type' => 'boundary');
}
// Otherwise, just return the single boundary character
if ($string[0] === '.' || $string[0] === ',') {
$type = $string[0];
} else {
$type = 'boundary';
}
return array('token' => $string[0], 'type' => $type);
}
// Whitespace
if (in_array($string[0], self::$whitespace)) {
for ($i = 1, $length = strlen($string); $i < $length; $i++) {
if (!in_array($string[$i], self::$whitespace)) {
break;
}
}
return array('token' => substr($string, 0, $i), 'type' => 'whitespace');
}
if (!self::$init) {
//Sort reserved word list from longest word to shortest
usort(self::$reserved, array('SqlFormatter', 'sortLength'));
//Combine boundary characters and whitespace
self::$all_boundaries = array_merge(self::$boundaries, self::$whitespace);
self::$init = true;
}
//a reserved word cannot be preceded by a '.'
//this makes it so in "mytable.from", "from" is not considered a reserved word
if (!$previous || !isset($previous['token']) || $previous['token'] !== '.') {
// Reserved word
$test = strtoupper($string);
foreach (self::$reserved as $word) {
$length = strlen($word);
if (substr($test, 0, $length) === $word) {
if (isset($string[$length]) && !in_array($string[$length], self::$all_boundaries)) {
continue;
}
if (in_array($word, self::$special_reserved)) {
$type = 'special reserved';
} else {
$type = 'reserved';
}
return array('token' => substr($string, 0, $length), 'type' => $type);
}
}
}
// Look for first word separator
for ($i = 1, $length = strlen($string); $i < $length; $i++) {
if (in_array($string[$i], self::$all_boundaries)) {
break;
}
}
$ret = substr($string, 0, $i);
if (is_numeric($ret)) {
$type = 'number';
} else {
$type = 'word';
}
return array('token' => $ret, 'type' => $type);
}
