Often times I find myself having to handle a particular portion of a SQL statement via a script. I've written a lot of specialized regular expressions over time to handle these tasks, but the one that I've had to write the most is a basic 'tokenizer' which understands the quoting semantics of a MySQL statement. That is, it understands `BackTick`, 'Single Quote' and "Double Quoted" strings.

#PHP Heredoc syntax
$regex = <<< END_OF_REGEX
/
  [^ \"'`(),]*\([^)]*\)    #match functions like concat(x,"y",`a`.`z`) or sum(`xyz`);
  |\([^)]*?\)              #match grouped items
  |"[^"]*?"                #match double quoted items
  |'[^']*?'                #match single quoted items
  |`[A-Za-z0-9_ .'\"()+\\-&^%\$+?%\\/\\\\!`]+`  #match backtick mysql names
  |[^ ,]+                  #match keywords, operators and aliases
  |,
/xi
END_OF_REGEX;

EDIT: After some



  [Read more...]