5

I'm plotting a curve from some data points:

points = [[0.0, 0.0], [1, 3], [3, 5]]

... and I want it to reach, say, y = 10 as x goes to infinity. In other words, the extra "point" to append would be: [float('inf'), 10]

Is there a way to tell matplotlib to plot the asymptotic behaviour of this last part of the curve?

Matplotlib disregards float('info'), so I guess there should be another way to do this.

Ricky Robinson
  • 21,798
  • 42
  • 129
  • 185
  • possible duplicate of [how to handle an asymptote/discontinuity with Matplotlib](http://stackoverflow.com/questions/2540294/how-to-handle-an-asymptote-discontinuity-with-matplotlib) – jonrsharpe Jan 09 '14 at 14:30
  • yes, I read it, but I didn't get any of the three answers there. How can I apply them to my problem? – Ricky Robinson Jan 09 '14 at 14:33
  • `matplotlib` won't plot infinite values (`np.inf` or `float('inf')`, because that doesn't make sense! You could try using a large but non-infinite number, e.g. `sys.maxsize`. – jonrsharpe Jan 09 '14 at 14:38
  • 1
    Analytically speaking of course it makes sense... maybe numerically it's not feasible. Ok. – Ricky Robinson Jan 09 '14 at 14:57
  • It don’t get the argument why it soesn’t make sense. It it wouldn’t make sense, `np.inf` wouldn’t be a valid floating point value – Frédéric Grosshans Oct 28 '21 at 10:16

1 Answers1

6

You can't plot a function of infinity because it isn't a number. I think the only way to achieve the effect that you want is with a broken axis and the data point plotted at some arbitrarily large number. An example of how to do the broken plot is available from the official matplotlib site.

The basic idea is to create two subplots with borders cleverly removed so as to look like a single plot then add some fancy diagonal lines to make it look like a conventional "broken" plot. You can plot your infinity data point in its own subplot by substituting a sufficiently large number (to make the curve look right). Then just relabel the tick to infinity (ax.set_xticks([large_number]) ax.set_xticklabels(['\infty']) assuming ax is the second plot that only has your infinity point) and you're in business!

I have recently used a modified form of the above example to plot a point "at infinity". I have code that does this but it's full of references to my specific data so it will probably be about as helpful to you as the official example.

mmdanziger
  • 4,466
  • 2
  • 31
  • 47