Search code examples
pythonmatplotlibsubplotaxes

Prob using axes for subplot an get AttributeError: 'numpy.ndarray' object has no attribute 'plot'

'''
matplotlib.use("Qt5Agg")

pas = [1,2,3,4,5,6,7,8,9,10]

# set up l'organisation de la figure
fig, axs = plt.subplots(nrows=3,ncols=len(pas))


manager = plt.get_current_fig_manager()
manager.window.showMaximized()
plt.show()
     
axs[1, 0+1].set(ylabel='fréquence (set test)')
for iii in range(0, len(pas)):
    
    bins = np.linspace(round(y[:,iii].min()), round(y[:,iii].max()), 20)
    axs[0, iii].hist(yhat_graph[:,iii], bins, color = 'steelblue', alpha = 0.5, label='simulation')
    axs[0, iii].hist(y[:,iii], bins, color = 'orange', alpha=0.5, label='ref. apport hist.')
    axs[0, iii].set(xlabel='apports m3/sec')

    axs[0, iii].legend(loc='upper right')
    axs[0, iii].set(title='prévision J+'  + str(iii+1) + ' (set test)' )
    

axs[1, 0].set(ylabel='Variations journalières Jn - Jn-1 (set test) \n observation')
for iii in range(0, len(pas)-1):
    axs[1, iii].plot()
    graph_scatter(ecart_valid_y[iii],
                  ecart_yhat[iii], True,'simulation','','var. j' + str(iii+1) +' - j ' + str(iii) ,'steelblue')
    axs[1, iii].set(xlabel='simulation')


if history == 'missing':
    print('pas de fichier history')
    axs[2,0:].plot()
else:
    axs[2, 0].plot()
    graph_loss(history)
    axs[2,1:].plot()

graph_sim_multiStep(y[-windowGraph[0]:-windowGraph[1]], yhat_graph[-windowGraph[0]:-windowGraph[1]], nash, kge, titre)

'''

with this line "axs[2,1:].plot()"

I've got this error: AttributeError: 'numpy.ndarray' object has no attribute 'plot'

the function 'graph_loss' and 'graph_scatter' work fine lonely

Solution

  • As you say, the problem is with this line:

    axs[2,1:].plot()

    In your code, axs is a 3x10 numpy array of AxesSubplot objects. I assume what you're trying to do is to call the plot() method on several of those objects at once, but numpy arrays don't support method calls like that. You could look at the vectorize() method, but I think it's clearest to just use a for loop. Here's a small example that calls plot() with some data on a couple of subplots, then calls plot() with no parameters on the rest of the subplots.

    import matplotlib.pyplot as plt

fig, axs = plt.subplots(nrows=2,ncols=3)

axs[0, 0].plot([1, 4, 2])
axs[0, 2].plot([4, 1, 2])

# Could probably remove the next three lines.
axs[0, 1].plot()
for iii in range(3):
    axs[1, iii].plot()

plt.show()

    I only wonder why you want to call plot() with no parameters in the first place. It just seems to change the axis scale a bit. Try removing those calls completely.