0

The time on the x-axis of my plot is in seconds since the epoch, and I want to convert them to "Year-Month-Day Hour:Minute" format. I follow the documentation here, and I get this traceback:

Traceback (most recent call last):
  File "/home/andreas/src/masiri/booking_algorythm/simple_booking.py", line 277, in <module>
    main()
  File "/home/andreas/src/masiri/booking_algorythm/simple_booking.py", line 273, in main
    accepted_list_of_bookings, accepted_bookings_a = process_booking_requests(randomized_list_of_bookings)
  File "/home/andreas/src/masiri/booking_algorythm/simple_booking.py", line 257, in process_booking_requests
    price = calculate_price_for_booking(trip_request, sequential_trip_requests,
  File "/home/andreas/src/masiri/booking_algorythm/simple_booking.py", line 226, in calculate_price_for_booking
    old_price = turn_booking_into_charging_plan(start_time, end_time, old_bookings_list)
  File "/home/andreas/src/masiri/booking_algorythm/simple_booking.py", line 209, in turn_booking_into_charging_plan
    plot_solution(sub_metric_v, soc_plot, charge_plot, discharge_plot, bookings_list, instability_ranges)
  File "/home/andreas/src/masiri/booking_algorythm/simple_booking.py", line 195, in plot_solution
    fig.autofmt_xdate()
  File "/home/andreas/src/masiri/venv/lib/python3.10/site-packages/matplotlib/figure.py", line 249, in autofmt_xdate
    for label in ax.get_xticklabels(which=which):
  File "/home/andreas/src/masiri/venv/lib/python3.10/site-packages/matplotlib/axes/_base.py", line 73, in wrapper
    return get_method(self)(*args, **kwargs)
  File "/home/andreas/src/masiri/venv/lib/python3.10/site-packages/matplotlib/axis.py", line 1381, in get_ticklabels
    return self.get_majorticklabels()
  File "/home/andreas/src/masiri/venv/lib/python3.10/site-packages/matplotlib/axis.py", line 1345, in get_majorticklabels
    self._update_ticks()
  File "/home/andreas/src/masiri/venv/lib/python3.10/site-packages/matplotlib/axis.py", line 1191, in _update_ticks
    major_labels = self.major.formatter.format_ticks(major_locs)
  File "/home/andreas/src/masiri/venv/lib/python3.10/site-packages/matplotlib/ticker.py", line 233, in format_ticks
    return [self(value, i) for i, value in enumerate(values)]
  File "/home/andreas/src/masiri/venv/lib/python3.10/site-packages/matplotlib/ticker.py", line 233, in <listcomp>
    return [self(value, i) for i, value in enumerate(values)]
  File "/home/andreas/src/masiri/venv/lib/python3.10/site-packages/matplotlib/dates.py", line 640, in __call__
    result = num2date(x, self.tz).strftime(self.fmt)
  File "/home/andreas/src/masiri/venv/lib/python3.10/site-packages/matplotlib/dates.py", line 533, in num2date
    return _from_ordinalf_np_vectorized(x, tz).tolist()
  File "/home/andreas/src/masiri/venv/lib/python3.10/site-packages/numpy/lib/function_base.py", line 2328, in __call__
    return self._vectorize_call(func=func, args=vargs)
  File "/home/andreas/src/masiri/venv/lib/python3.10/site-packages/numpy/lib/function_base.py", line 2411, in _vectorize_call
    outputs = ufunc(*inputs)
  File "/home/andreas/src/masiri/venv/lib/python3.10/site-packages/matplotlib/dates.py", line 354, in _from_ordinalf
    np.timedelta64(int(np.round(x * MUSECONDS_PER_DAY)), 'us'))
OverflowError: int too big to convert

This is my code (I will supply a minimal example, but as far as I can see, this is purely time conversion related).

def plot_solution(data, soc_plot,charge_plot, discharge_plot, bookings=None, boundaries=None):
    my_fmt = mdates.DateFormatter('%Y-%m-%d %H:%M')
    fig, ax_left = plt.subplots()
    ax_right = ax_left.twinx()
    ax_left.xaxis.set_major_formatter(my_fmt)
    ax_right.xaxis.set_major_formatter(my_fmt)
    time_axis = data["time"]

    ax_right.plot(time_axis, soc_plot, label="State of Charge")
    ax_left.plot(time_axis, data["price"], label="price (€/kWs)")
    ax_left.plot(time_axis, data["price_co2"], label="price including CO_2 costs (€/kWs)")
    #ax_right.plot(data["time"], charge_plot, label="charge")
    #ax_right.plot(data["time"], discharge_plot, label="discharge")
    bookings_flag = True
    if bookings is not None:
        for booking in bookings:
            if bookings_flag:
                ax_left.axvspan(booking["start"], booking["start"] + booking["duration"], facecolor='0.4', alpha=0.5, label="booking")
                bookings_flag = False
            else:
                ax_left.axvspan(booking["start"], booking["start"] + booking["duration"], facecolor='0.4', alpha=0.5, label="booking")
    boundaries_flag = True
    if boundaries is not None:
        for boundary in boundaries:
            if boundaries_flag:
                ax_left.axvspan(time_axis[boundary["start"]], time_axis[boundary["end"]], facecolor='0.9', alpha=0.5, label="boundary")
                boundaries_flag = False
            else:
                ax_left.axvspan(time_axis[boundary["start"]], time_axis[boundary["end"]], facecolor='0.9', alpha=0.5)
    ax_left.set_xlabel("time in seconds")
    ax_left.set_ylabel("price in €/kWs", color="blue")
    ax_right.set_ylabel("State of Charge (SoC), normalized", color="black")
    legend_1 = ax_left.legend(loc=2, borderaxespad=1.)
    legend_1.remove()
    ax_right.legend(loc=1, borderaxespad=1.)
    ax_right.add_artist(legend_1)
    fig.autofmt_xdate()
    plt.show(block=True)

Here is an example of my time_axis data:

>>> time_axis
array([1635905700, 1635906600, 1635907500, 1635908400, 1635909300,
       1635910200, 1635911100, 1635912000, 1635912900, 1635913800,
Andreas Schuldei
  • 343
  • 1
  • 15
  • Please give a few examples of what is in data[‘time’]. You say it’s seconds since the epoch but don’t show how you convert it to a date that Matplotlib will accept. Just running autofmt_date won’t convert a list of floats to meaningful dates – Jody Klymak Feb 25 '23 at 15:46
  • @JodyKlymak done. Perhaps that is the issue already? I was under the impression that the conversion was done by matplotlib. – Andreas Schuldei Feb 25 '23 at 15:53

1 Answers1

1

Dates in Matplotlib are stored as days since 1970 (if using Matplotlb >=3.5, the epoch was 0000 before). So if you want to use something like plot_date on straight floats you can if you convert the floats from seconds to days. However, better is probably to convert to datetime64.

import matplotlib.pyplot as plt
import numpy as np
import matplotlib as mpl

time = np.array([1635905700, 1635906600, 1635907500, 1635908400, 1635909300,
       1635910200, 1635911100, 1635912000, 1635912900, 1635913800,])
y = np.arange(10)

plt.rcParams['date.converter'] = 'concise'

fig, ax = plt.subplots(2, 1, layout='constrained')
ax[0].plot(time/24/3600, y)
ax[0].xaxis_date()

ax[1].plot(time.astype('datetime64[s]'), y)

plt.show()

Note that I've also used the concise date converter https://matplotlib.org/stable/gallery/ticks/date_concise_formatter.html but that is just aesthetics.

xaxis_date (or ax.plot_date or fig.autofmt_xdate) are just short forms to set the units on the xaxis to datetime.

enter image description here

Jody Klymak
  • 4,979
  • 2
  • 15
  • 31