Ogden, Craig N wrote:
> Those milliseconds are very important to my data.
Then the day isn't. For the simple reason that there's no way you'll be
able to resolve a day's worth of millisecond-precision data on a single
plot --- no output device has that many pixels in a row.
> The only problem is
> that gnuplot, as far as I can tell, doesn't do milliseconds when it
> comes to time.
The actual problem is that the C runtime library doesn't do sub-second
time. gnuplot's functions are a direct equivalent of ISO and POSIX
standard <time.h> functions, with just two small twists: the epoch is in
Y2K, and our time_t is a double.
> Is there any way around it the millisecond problem?
Throw away everything above minutes, and use a format string:
plot 'data' u (3600*$1+60*$2+$3):4 "%*lf/%*lf:%lf:%lf:%lf %lf"
Or, if you're feeling brave, patch your gnuplot. There's a patch to
extend %S to take floating-point input at our SourceForge site, IIRC.
|