> > I would love to do an sql query that would look like the following > ones. I wonder if they are possible and valid applications for SQL and > what would be the proper implementation for these. I know I can calculate > all this using C, but it would be most useful (to my later projects) to do > it in SQL directly (my last example gives you an idea where I'm going). > > select max(opening_price, closing_price, high_price) - min(opening_price, > closing_price, low_price) as day_range... > select closing_price, moving_average(20,closing_price), > exp_mov_avg(20,closing_price)... > select closing_price, moving_average( funky_oscillator( closing_price ) > )... >
I have described the problem that I was facing. You guys provided all the pieces of the puzzle. I now have to solve the problem. Here's the path that I have decided to take for now, more or less in priority: - Implement what I need now in plain C++, after the data is extracted from sqlite. - Understand what analytical functions are, look at their implementations and attempt my own in plain C++ - Experiment at creating my own sqlite functions, agg.functions and virtual tables as these will be instrumental to my goal or future goals. - Look into the libraries you guys suggested for analyzing the data (I keep this mostly last because my analytic skills are still at a minimum, so this could be the most difficult path and I also see educational value in re-inventing the wheel for now, but I do understand this step may be required to move to higher levels of analysis). - Look at sqlite's source code and try to implement analytical functions in a way that leads to an optimization better than log(n^2) and contribute my findings on this topic back to the community. Thank you all very much for all your answers, they have been most useful. Simon _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
