Remove plot from matplotlib subplot, but keep legend visible-CodePudding

I have a very customized subplot set up.

fig = plt.figure(figsize=(12, 10))
gs = fig.add_gridspec(nrows=2, ncols=2, width_ratios=[3, 1])
ax = fig.add_subplot(gs[:, 0])
ax3 = fig.add_subplot(gs[-1, -1])
ax4=fig.add_subplot(gs[0, 1])

This sets up 3 slots for plotting: one that takes up half the space on the left, and two smaller ones on the right. However, I only want the bottom right to actually be a plot. I want the top right to be the space where the legend for the larger plot on the left to go. I could just use the axes from ax to do this, but that shifts the whole plotting space off. Instead I thought of trying to just create ax4 and place the ax legend there.

lines = []
labels = []

for ax in fig.get_axes():
    ln, la = ax.get_legend_handles_labels()
    lines.extend(ln)
    labels.extend(la)
legend = ax4.legend(lines, labels, labelspacing=0.1, loc=(-0.3,0.6), fontsize='xx-large')
fig.tight_layout()

This puts the legend exactly where I want it, but the blank figure shows up, which I don't want. Is it possible to accomplish what I want using this method? If not, what is my alternative? Picture below to better understand.

CodePudding user response：

You can use ax4.axis('off') to make axis 4 invisible if you want to stick to your approach. However, I don't see why you don't just skip creating axis 4 and just use fig.legend() instead of ax.legend(). Then the legend is placed outside the axis and you can then control the exact position just as you already did with the loc keyword.