Change marker in the legend in matplotlib


Suppose that you plot a set of data:

plt.plot(x,y, marker='.', label='something')

On the display, you will obtain . something, but how do you do to change it to - something, so that the marker that appears in the legend is a line a not a dot?

Asked By: user2820579



The solution surely depends on the criterion by which you want to transform the marker. Doing this manually is straight forward:

import matplotlib.pyplot as plt

line, = plt.plot([1,3,2], marker='o', label='something')
plt.legend(handles = [plt.plot([],ls="-", color=line.get_color())[0]],

enter image description here

Doing the same in an automated way, i.e. each line in the plot gets its corresponding legend handle, which is a line in the same color, but without markers:

import matplotlib.pyplot as plt
from matplotlib.legend_handler import HandlerLine2D

plt.plot([1,3,2], marker='o', label='something')
plt.plot([2,3,3], marker='o', label='something else')

def update_prop(handle, orig):


enter image description here

To follow up with more info. If you want to make your marker in the legend more visible when using plt.scatter(), you can do the following to automate the marker change process.

import matplotlib.pyplot as plt
import matplotlib.patches as mpatches

handles, labels, plt.gca().get_legend_handles_labels()
updated_handles = []
for handle in handles:
  updated_handles.append(mpatches.Patch(color=handle.get_facecolor(), label=handle.get_label()))
by_label = dict(sorted(dict(zip(all_labels, updated_handles)).items()))
plt.figlegend(by_label.values(), by_label.keys(), ...)

If you would like to use a specific axis, change plt.gca() with yours. Also, on line 8 I sorted the legend, which is something you can choose not to do.

Answered By: user12259798

It works for 3d plots slightly differently:

import matplotlib.pyplot as plt
import matplotlib.patches as mpatches

ax = plt.gca()
handles, labels =  ax.get_legend_handles_labels()
updated_handles = []

for handle in handles:              
by_label = dict(sorted(dict(zip(labels,
ax.legend(by_label.values(), by_label.keys())
Answered By: msch
Categories: questions Tags: ,
Answers are sorted by their score. The answer accepted by the question owner as the best is marked with
at the top-right corner.