How do I set color to Rectangle in Matplotlib?

Turns out, you need to do ax.add_artist(Rectangle) to have the color specifications work; when using patches.append(Rectangle), the rectangle is shown in blue (on my PC, at least) ignoring any color specification.

Btw, note that artists — Matplotlib 1.2.1 documentation: class matplotlib.patches.Rectangle states that there is

• edgecolor - for stroke color
• facecolor - for fill color

... and then there is color - which basically sets both stroke and fill color at the same time.

Here is the modified OP code, which I've tested on Linux (Ubuntu 11.04), python 2.7, matplotlib 0.99.3:

import matplotlib.pyplot as plt
import matplotlib.collections as collections
import matplotlib.ticker as ticker

import matplotlib
print matplotlib.__version__ # 0.99.3

fig=plt.figure() #pylab.figure()
ax=fig.add_subplot(111)

ax.set_xlim([-400, -380]) #pylab.xlim([-400, 400])
ax.set_ylim([-400, -380]) #pylab.ylim([-400, 400])
patches = []
polygon = plt.Rectangle((-400, -400), 10, 10, color='yellow') #Rectangle((-400, -400), 10, 10, color='y')
patches.append(polygon)

pol2 = plt.Rectangle((-390, -390), 10, 10, facecolor='yellow', edgecolor='violet', linewidth=2.0)
ax.add_artist(pol2)


p = collections.PatchCollection(patches) #, cmap=matplotlib.cm.jet)
ax.add_collection(p)
ax.xaxis.set_major_locator(ticker.MultipleLocator(20)) # (MultipleLocator(20)) 
ax.yaxis.set_major_locator(ticker.MultipleLocator(20)) # (MultipleLocator(20)) 

plt.show() #pylab.show()

this is the output:

To avoid calling .add_patch() multiple times (often the purpose of using PatchCollection in the first place), you can pass a ListedColormap to the PatchCollection via cmap=.

This looks as follows (modified from fraxel's answer):

import matplotlib
import matplotlib.pyplot as plt
from matplotlib.colors import ListedColormap
from matplotlib.collections import PatchCollection
import numpy as np

fig = plt.figure()
ax = fig.add_subplot(111)
patches_list = []
color_list = []
patches_list.append(matplotlib.patches.Rectangle((-200,-100), 400, 200))
color_list.append('yellow')
patches_list.append(matplotlib.patches.Rectangle((0,150), 300, 20))
color_list.append('red')
patches_list.append(matplotlib.patches.Rectangle((-300,-50), 40, 200))
color_list.append('#0099FF')
patches_list.append(matplotlib.patches.Circle((-200,-250), radius=90))
color_list.append('#EB70AA')

our_cmap = ListedColormap(color_list)
patches_collection = PatchCollection(patches_list, cmap=our_cmap)
patches_collection.set_array(np.arange(len(patches_list)))
ax.add_collection(patches_collection)

plt.xlim([-400, 400])
plt.ylim([-400, 400])
plt.show()

Result: cmap_approach_result

I ran into this same issue. Passing in color= to Rectangle() will only work if you specify match_original=True when you create your PatchCollection(). See PatchCollection() for the details.

A simplified version of @Didillysquat's example is

import matplotlib.pyplot as plt
from matplotlib.collections import PatchCollection

fig = plt.figure()
ax = fig.add_subplot(111)
patches_list = []
color_list = []
patches_list.append(matplotlib.patches.Rectangle((-200,-100), 400, 200, color='yellow'))
patches_list.append(matplotlib.patches.Rectangle((0,150), 300, 20, color='red'))
patches_list.append(matplotlib.patches.Rectangle((-300,-50), 40, 200, color='#0099FF'))
patches_list.append(matplotlib.patches.Circle((-200,-250), radius=90, color='#EB70AA'))

# Make sure you use match_original=True
patches_collection = PatchCollection(patches_list, match_original=True)
ax.add_collection(patches_collection)

plt.xlim([-400, 400])
plt.ylim([-400, 400])
plt.show()