question

There are already a number of tools out there that do this. You might want to take a look at them. There's [SSMS Tools Pack][1], [TSQL Prettifier][2], and my company's product, [SQL Prompt][3]. I think there are a couple of others as well. The first two in the list are free. [1]: http://www.ssmstoolspack.com/ [2]: http://extras.sqlservercentral.com/prettifier/prettifier.aspx [3]: http://www.red-gate.com/products/sql-development/sql-prompt/

I am not sure why you would want to attempt such a complex task but, if you are still going to do it anyways, then the starting point would be to parse the given SQL and validate that it is correct and then tokenize it into keywords. I did a quick google search and found the following links: [Google][1] [CodeProject][2] [Devioblog][3] [1]: http://www.google.co.in/#sclient=psy-ab&hl=en&source=hp&q=sql+parser+c%23&pbx=1&oq=sql+parser+c&aq=1&aqi=g2g-v2&aql=&gs_sm=3&gs_upl=0l0l1l375l0l0l0l0l0l0l0l0ll0l0&bav=on.2,or.r_gc.r_pw.,cf.osb&fp=ddbe45f2069e114c&biw=1440&bih=719 [2]: http://www.codeproject.com/Articles/32524/SQL-Parser [3]: http://devio.wordpress.com/2009/06/29/parsing-sql-into-xml-and-c-objects/