Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.graphics.apps.gnuplot > #3974

Re: consistency of function .vs. data plotting

Show all headers | View raw

Am 19.06.2018 um 18:41 schrieb palme@kapott.org:

> But the approach using the pseudo data file plots also points for the complex results outside the [-1,1] range.
> 
> See yourself:
>   print "using pure function"
>   plot asin(x)
>   pause -1 "press enter"
>   print "using pseudo data file"
>   plot '+' using 1:(asin($1)) with lines
>   pause -1 "press enter"
> 
> So I have two questions:
> 
> 1) Where exactly is my misunderstanding of the documentation, regarding the way how data file plotting works in comparison to plotting pure functions?
> 
> 2) How can I avoid these complex results to be plotted with the second approach?


gnuplot does not internally report complex values as _undefined_ in data
plots, but does so in function plots. That's imo just a bug.

Of course a data plot containing non-real values makes no real sense,
and you can't even directly read in complex values from a data file,
because there is no predefined format for them.

   set sample 11
   set table
   plot "+" us 1:(1/$1)  # 1/0 _is_ reported as undefined
   plot "+" us 1:(sqrt($1))  # sqrt of negative values isn't

As a workaround, you could do sth like following

   f(x) = ...
   plot "+" us 1:(imag(f(x) =0 ? f(x) : NaN)

Back to comp.graphics.apps.gnuplot | Previous | Next — Previous in thread | Next in thread | Find similar

Thread

consistency of function .vs. data plotting palme@kapott.org - 2018-06-19 09:41 -0700
  Re: consistency of function .vs. data plotting Karl Ratzsch <mail.kfr@gmx.net> - 2018-06-25 16:06 +0200
    Re: consistency of function .vs. data plotting Karl Ratzsch <mail.kfr@gmx.net> - 2018-06-25 23:55 +0200
  Re: consistency of function .vs. data plotting palme@kapott.org - 2018-06-25 08:43 -0700

csiph-web