Garbled title string in plt.set

I have an xarray.Dataset, ds, with a time variable that is:

array([cftime.DatetimeGregorian(2021, 10, 16, 9, 50, 1, 0)], dtype=object)

I would like to use the date in the plot title, as:

fig = plt.figure()
ax = plt.axes(projection = ccrs.PlateCarree(central_longitude=200))  # Orthographic
day_str = np.datetime_as_string(ds.time, unit='D').tobytes().decode()
ax.set_title(day_str, size = 10.)

The problem I'm having is that, when I use the variable day_str in the title it is all garbled.

type(day_str) returns str. When I type print(day_str), I get: 2021-10-16, as expected. So, I do not think this is a cftime to python datetime issue. What am I missing?

Another thing that is relevant is that evaluating 'foo' day_str, gives:

'foo2\x00\x00\x000\x00\x00\x002\x00\x00\x001\x00\x00\x00-\x00\x00\x001\x00\x00\x000\x00\x00\x00-\x00\x00\x001\x00\x00\x006\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00'

EDIT: Well, this isn't very elegant, but it works:

fig = plt.figure()
ax = plt.axes(projection = ccrs.PlateCarree(central_longitude=200))  # Orthographic
daystr=ds.time.dt.day.astype(str).values[0]
monstr=ds.time.dt.month.astype(str).values[0]
yrstr=ds.time.dt.year.astype(str).values[0]
day_str = monstr '-' daystr '-' yrstr
ax.set_title('SST, '   day_str, size = 10.)

CodePudding user response：

As far as I interpret your question, this is what I get when I convert your DatetimeGregorian objet to a string:

import cftime
str(cftime.DatetimeGregorian(2021, 10, 16, 9, 50, 1, 0))

Resulting string:

'2021-10-16 09:50:01'

CodePudding user response：

The section on time series in the xarray documentation is a good resource: http://xarray.pydata.org/en/stable/user-guide/time-series.html

A good way to handle this would be to use the using the special .dt accessor with the method .strftime:

day_str = ds.time.isel(time=-1).dt.strftime("%a, %b %d %H:%M").values

or, equivalently in this case,

day_str=ds.time.dt.strftime("%a, %b %d %H:%M").values[0]

With this approach, 'SST, ' day_str gives

'SST, Tue, Oct 19 09:50'