strange behavior of matplotlib plot on numpy matrix element

Question

I'm using python3.X.

I came across some very strange behavior in matplotlib on a numpy matrix element as an ilustration I wanted to plot a simple sinc() function:

import numpy as np
import matplotlib.pyplot as plt

t=np.matrix(np.linspace(-10,10,1024))
x=np.sinc(t)

plt.plot(t,x,color='blue', linestyle='solid', linewidth=2)
plt.show()

The above piece of code generates:

While replacing the plt.plot(...) with:

plt.plot(t,x,'-ob')

generates:

I could not figure out the reason for this behavior, would appreciate some help

In your 1st case plt.plot is interpreting your input as 1024 data sets and plotting each of them individually. Replace with t.T and x.T ( transpose your arrays) and it works fine — Brenlla
– Brenlla, Commented Sep 19, 2019 at 9:58
The same is true for your 2nd case, but since you are using dots rather than lines the output looks Ok. — Brenlla
– Brenlla, Commented Sep 19, 2019 at 9:59
@Brenlla can you post it as answer, for me to accept, Thanks — David
– David, Commented Sep 19, 2019 at 10:00

Brenlla · Accepted Answer · 2019-09-19 10:10:04Z

2

According to the docs, when using 2-D arrays, plot interprets the columns as separate data sets. Therefore, in your 1st case you are plotting 1024 lines with one point each. Since line plots work by drawing lines between points, nothing is being displayed:

t=np.matrix(np.linspace(-10,10,1024))
x=np.sinc(t)

plt.plot(t,x,color='blue', linestyle='solid', linewidth=2)
# plot shows nothing

Transpose your arrays to single column and it works fine:

plt.plot(t.T,x.T,color='blue', linestyle='solid', linewidth=2)
# plot shows line

The 2nd case works fine because when plotting dots, one is drawn for each data point. It is probably still faster to just plot one data set though:

plt.plot(t.T,x.T,'-ob')
# same output, probably faster

answered Sep 19, 2019 at 10:10

Brenlla

1,4811 gold badge11 silver badges24 bronze badges

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

3 Comments

David Over a year ago

Another question, why when i change t to np.array it works without the transpose; what is the essential difference between the 2 types?; Also would appreciate an upvote

Brenlla Over a year ago

It is recommended you use array rather than matrix. Matrices are always 2-D, whereas arrays can have any dimensionality. For this case, if your create 1-D arrays (check t.ndim), transposing does nothing since there is only one dimension.

David Over a year ago

I check the ndim and as you said, indeed for the np.array its just 1D while for the np.matrix is 2D; Thanks for pointing the problem

Collectives™ on Stack Overflow

strange behavior of matplotlib plot on numpy matrix element

1 Answer 1

3 Comments

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

3 Comments

Related