10

I'm trying to plot a subset of some data, but the y-axis limits are not updated properly after I set the x-axis limits. Is there a way to have matplotlib update the y-axis limits after setting the x-axis limits?

For example, consider the following plot:

import numpy
import pylab
pylab.plot(numpy.arange(100)**2.0)

which gives: Plot 1 full range

which works fine. However if I want to only view the part from x=0 to x=10, the y-scaling is messed up:

pylab.plot(numpy.arange(100)**2.0)
pylab.xlim(0,10)

which gives: Plot 1 subset.

In the former case, the x- and y-axis are scaled properly, in the latter case, the y-axis is still scaled the same, even if the data is not plotted. How do I tell matplotlib to update the y-axis scaling?

Obvious workarounds would be to plot a subset of the data itself, or to reset the y-axis limits manually by inspecting the data, but those are both rather cumbersome.

Update:

The example above is simplified, in the more general case one has:

pylab.plot(xdata, ydata1)
pylab.plot(xdata, ydata2)
pylab.plot(xdata, ydata3)
pylab.xlim(xmin, xmax)

Setting the y-axis range manually is of course possible

subidx = N.argwhere((xdata >= xmin) & (xdata <= xmax))
ymin = N.min(ydata1[subidx], ydata2[subidx], ydata3[subidx])
ymax = N.max(ydata1[subidx], ydata2[subidx], ydata3[subidx])
pylab.xlim(xmin, xmax)

but this is cumbersome to say the least (imho). Is there a faster way to do this without manually calculating the plotranges? Thanks!

Update 2:

The function autoscale does some scaling and seems the right candidate for this job, but treats the axes independently and only scales to the full data range, no matter what the axes limits are.

Improve this question
2

1 Answer 1

Reset to default
2

In what sense do you mean inspecting the data is cumbersome? If in terms of writing code, then it's not so bad. Try something like

pylab.ylim(numpy.min(data), numpy.max(data))

...where data can be numpy.arange(100)[0:11].

In the general case, if you have xdata and ydata (but supposing they are sorted) you would have to so something like

from bisect import bisect
sub_ydata = ydata[bisect(xdata, xmin):bisect(xdata, xmax)]
pylab.ylim(numpy.min(sub_ydata), numpy.max(sub_ydata))

If you mean that it's a hard thing computationally, then I don't really see how matplotlib could perform it without such calculations.

Improve this answer
Sign up to request clarification or add additional context in comments.

5 Comments

I didn't mean it's computationally hard, I just wonder if/how matplotlib can update the axis by itself. Obviously matplotlib gets the axis range correctly if you plot the full range of the data, but it doesn't act so logical when you update one of the axis ranges, imho. Also in the general case when you have xdata and ydata, setting things manually becomes more cumbersome.
The logic is that the automatic scaling stops once you adjust the limits manually. It makes sense; consider how annoying it would be if for some reason you did not want matplotlib to adjust automatically, but it would. The general case is not much worse, let me add a short example of what I mean.
I agree that automatic scaling could interfere with manual scaling, but IMHO this is not the case here. Especially if I set the xrange before plotting, I would expect matplotlib to change the yrange accordingly. Gnuplot does this, for example. In any case I assume the answer to my question is 'No', so thanks :)
Ah, I found it, autoscale seems to do what I want!
No wait, autoscale() also ignores any axes limits and only looks at the data given to plot(). This would've been a good candidate to solve this issue but it does not, alas.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.