On Wed, Apr 04, 2012 at 12:05:44PM +0200, Steinar Midtskogen scratched on the wall: > [Jean-Christophe Deschamps] > > > You're going to have at most one random sample in every slice of 320 > > s. The GROUP BY clause will select only one for you and the query can > > be as simple as: > > > > select sample from from mytable group by timestamp / 320 order by > > timestamp; > > Ah. I didn't think of that. It's even better than getting every nth > row, since I get one sample for a fixed period, which is what I really > want. And yet better, I suppose I could do something like SELECT > min(sample), max(sample) FROM mytable GROUP BY timestamp / 3600 and > use financebars or similar in gnuplot to avoid missing the extremes in > the plot, making it appear more or less identical as if I had plotted > every value.
Not to mention avg(). You might want to have a look at how RRDtool deals with condensing data. It is common pratice to plot average, min, and max to preserve outliers, while still showing trends. http://oss.oetiker.ch/rrdtool/ -j -- Jay A. Kreibich < J A Y @ K R E I B I.C H > "Intelligence is like underwear: it is important that you have it, but showing it to the wrong people has the tendency to make them feel uncomfortable." -- Angela Johnson _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users