使用日期作为 x 轴的不断增长的 matplotlib 条形图

Question

I have some code that should, in theory, save an image of a bar chart that grows by thirty days each time a loop completes, here is the code:

import matplotlib.pyplot as plt 
import datetime as dt
import matplotlib.dates as mdates

f, ax = plt.subplots()
ax.xaxis.set_major_locator(mdates.YearLocator(1))
ax.xaxis_date()

xi = dt.datetime(1900,1,1)
xf = dt.datetime(1910,1,1)

x = xi
num = 0

while x < xf:
    #add 30 days to the current date
    x += dt.timedelta(30)

    #tranform datetime data into matplotlib number dates
    xmin = mdates.date2num(xi)
    xmax = mdates.date2num(x)

    #plot the bar graph that should grow by 30 days each loop
    ax.broken_barh([(xmin,xmax-xmin)], (0,1))

    #save the figure as a frame
    plt.savefig("frame-{}.png".format(num))
    num += 1

The problem is that I get a AttributeError: 'numpy.float64' object has no attribute 'toordinal' error every time I try to run it. I use this exact same code to create a plot that does not grow outside of a loop and it works. When I remove the ax.xaxis_date() portion of the code it works, except the x-axis of the chart is not formatted as a date.

I have another place where I use similar code, basically the same without the loop or the growing variable, and it works even with the ax.xaxis_date() function.

I've been troubleshooting this for a really long time and can't seem to figure it out, any help would be much appreciated. Thanks!

Answer 1

I have run your code as-is, simple copy-paste and it worked for me. It produced files named frame-*.png, * going from 0 to 121. I attach the first and final files below.

I am using Python 3.6.8 on Ubuntu 18.04.3 LTS with matplotlib 3.1.2.