has someone started writing a database engine?

Yes, but py pieces, and not yet in FASM at all !!

http://www.geocities.com/remyvincent/BinTree.zip
1. Somewhere in this package, parenthesis levels ARE HANDLED...
2. In the same package, Binary Tree are handled, but not with true evaluation...

http://www.geocities.com/remyvincent/9ParsersInAJS.zip
3. An SQL Parser SHOULD LOOK moreless like one of this 9 parsers (Pascal, C, Ada,QBasic,...)

4. I have somewhere (not downloadable) ".CSV" routines... may be CommaSeparated databases

So let's say the work is 15%~20% ADVANCED...., but in various languages, and not yet in FASM


And also, as it is big program, It would be nice from all of you if 3 or 4 teams would create their own "database engine" in FASM, and post the result here, so I would not be all alone... with this big project done by me alone... It's always hard to handle "big & proud" projects...

http://www.sqlite.org/opcode.html

SQLite already implements queries as opcodes (but the interface is normal SQL statements).

As Tomasz (or someone else ) proposed, maybe rewriting SQLite in ASM and making an interface to bypass the parser to send opcodes queries directly can be a good idea. More planification is needed though, because it could be great support object oriented databases too (which I think it's more interesting from the assembly point of view).