有时您会有一个子图网格，并且您希望有一个图例来描述每个子图的所有线条，如下图所示。
为此，您需要为图形创建全局图例，而不是在轴级别创建图例（这将为每个子图创建单独的图例）。这是通过调用实现的，如以下代码的代码所示。fig.legend()
fig, (ax1, ax2, ax3) = plt.subplots(1, 3, figsize=(10,4))
fig.suptitle('Example of a Single Legend Shared Across Multiple Subplots')
# The data
x = [1, 2, 3]
y1 = [1, 2, 3]
y2 = [3, 1, 3]
y3 = [1, 3, 1]
y4 = [2, 2, 3]
# Labels to use in the legend for each line
line_labels = ["Line A", "Line B", "Line C", "Line D"]
# Create the sub-plots, assigning a different color for each line.
# Also store the line objects created
l1 = ax1.plot(x, y1, color="red")[0]
l2 = ax2.plot(x, y2, color="green")[0]
l3 = ax3.plot(x, y3, color="blue")[0]
l4 = ax3.plot(x, y4, color="orange")[0] # A second line in the third subplot
# Create the legend
fig.legend([l1, l2, l3, l4], # The line objects
labels=line_labels, # The labels for each line
loc="center right", # Position of legend
borderaxespad=0.1, # Small spacing around legend box
title="Legend Title" # Title for the legend
)
# Adjust the scaling factor to fit your legend text completely outside the plot
# (smaller value results in more space being made for the legend)
plt.subplots_adjust(right=0.85)
plt.show()关于上述示例，需要注意以下几点：
l1 = ax1.plot(x, y1, color="red")[0]
当plot()被调用时，它返回一个line2D对象列表。在这种情况下，它只返回一个包含一个line2D对象的列表，该对象是通过[0]索引提取的，并存储在l1.
我们有兴趣包括在图例中的所有line2D对象的列表需要作为第一个参数传递给。的第二个参数也是必要的。它应该是一个字符串列表，用作图例中每一行的标签。fig.legend()fig.legend()
传递给的其他参数纯粹是可选的，只是帮助微调图例的美感。fig.legend()