df.plot() does not work for time series after deregistering converters Prophet

[sudhNau]

Code Sample, a copy-pastable example if possible

import pandas as pd
import numpy as np

from pandas.plotting import deregister_matplotlib_converters

deregister_matplotlib_converters()


ts = pd.Series(np.random.randn(1000), index=pd.date_range('1/1/2000', periods=1000))
ts.plot()

~/sandbox/pandas/pandas/plotting/_core.py in __call__(self, kind, ax, figsize, use_index, title, grid, legend, style, logx, logy, loglog, xticks, yticks, xlim, ylim, rot, fontsize, colormap, table, yerr, xerr, label, secondary_y, **kwds)
    808                            colormap=colormap, table=table, yerr=yerr,
    809                            xerr=xerr, label=label, secondary_y=secondary_y,
--> 810                            **kwds)
    811     __call__.__doc__ = plot_series.__doc__
    812

~/sandbox/pandas/pandas/plotting/_core.py in plot_series(data, kind, ax, figsize, use_index, title, grid, legend, style, logx, logy, loglog, xticks, yticks, xlim, ylim, rot, fontsize, colormap, table, yerr, xerr, label, secondary_y, **kwds)
    766                  yerr=yerr, xerr=xerr,
    767                  label=label, secondary_y=secondary_y,
--> 768                  **kwds)
    769
    770

~/sandbox/pandas/pandas/plotting/_core.py in _plot(data, x, y, subplots, ax, kind, **kwds)
    714         plot_obj = klass(data, subplots=subplots, ax=ax, kind=kind, **kwds)
    715
--> 716     plot_obj.generate()
    717     plot_obj.draw()
    718     return plot_obj.result

~/sandbox/pandas/pandas/plotting/_matplotlib/core.py in generate(self)
    214         self._compute_plot_data()
    215         self._setup_subplots()
--> 216         self._make_plot()
    217         self._add_table()
    218         self._make_legend()

~/sandbox/pandas/pandas/plotting/_matplotlib/core.py in _make_plot(self)
    987                              stacking_id=stacking_id,
    988                              is_errorbar=is_errorbar,
--> 989                              **kwds)
    990             self._add_legend_handle(newlines[0], label, index=i)
    991

~/sandbox/pandas/pandas/plotting/_matplotlib/core.py in _ts_plot(cls, ax, x, data, style, **kwds)
   1025         ax._plot_data.append((data, cls._kind, kwds))
   1026
-> 1027         lines = cls._plot(ax, data.index, data.values, style=style, **kwds)
   1028         # set date formatter, locators and rescale limits
   1029         format_dateaxis(ax, ax.freq, data.index)

~/sandbox/pandas/pandas/plotting/_matplotlib/core.py in _plot(cls, ax, x, y, style, column_num, stacking_id, **kwds)
   1002             cls._initialize_stacker(ax, stacking_id, len(y))
   1003         y_values = cls._get_stacked_values(ax, stacking_id, y, kwds['label'])
-> 1004         lines = MPLPlot._plot(ax, x, y_values, style=style, **kwds)
   1005         cls._update_stacker(ax, stacking_id, y)
   1006         return lines

~/sandbox/pandas/pandas/plotting/_matplotlib/core.py in _plot(cls, ax, x, y, style, is_errorbar, **kwds)
    581             else:
    582                 args = (x, y)
--> 583             return ax.plot(*args, **kwds)
    584
    585     def _get_index_name(self):

~/Envs/pandas-dev/lib/python3.7/site-packages/matplotlib/axes/_axes.py in plot(self, scalex, scaley, data, *args, **kwargs)
   1666         lines = [*self._get_lines(*args, data=data, **kwargs)]
   1667         for line in lines:
-> 1668             self.add_line(line)
   1669         self.autoscale_view(scalex=scalex, scaley=scaley)
   1670         return lines

~/Envs/pandas-dev/lib/python3.7/site-packages/matplotlib/axes/_base.py in add_line(self, line)
   1898             line.set_clip_path(self.patch)
   1899
-> 1900         self._update_line_limits(line)
   1901         if not line.get_label():
   1902             line.set_label('_line%d' % len(self.lines))

~/Envs/pandas-dev/lib/python3.7/site-packages/matplotlib/axes/_base.py in _update_line_limits(self, line)
   1920         Figures out the data limit of the given line, updating self.dataLim.
   1921         """
-> 1922         path = line.get_path()
   1923         if path.vertices.size == 0:
   1924             return

~/Envs/pandas-dev/lib/python3.7/site-packages/matplotlib/lines.py in get_path(self)
   1025         """
   1026         if self._invalidy or self._invalidx:
-> 1027             self.recache()
   1028         return self._path
   1029

~/Envs/pandas-dev/lib/python3.7/site-packages/matplotlib/lines.py in recache(self, always)
    668         if always or self._invalidx:
    669             xconv = self.convert_xunits(self._xorig)
--> 670             x = _to_unmasked_float_array(xconv).ravel()
    671         else:
    672             x = self._x

~/Envs/pandas-dev/lib/python3.7/site-packages/matplotlib/cbook/__init__.py in _to_unmasked_float_array(x)
   1388         return np.ma.asarray(x, float).filled(np.nan)
   1389     else:
-> 1390         return np.asarray(x, float)
   1391
   1392

~/Envs/pandas-dev/lib/python3.7/site-packages/numpy/core/numeric.py in asarray(a, dtype, order)
    536
    537     """
--> 538     return array(a, dtype, copy=False, order=order)
    539
    540

TypeError: float() argument must be a string or a number, not 'Period'

Problem description

Above code is the first time series plotting example in the docs. It throws a type error as follows

[paste the output of pd.show_versions() here below this line]

INSTALLED VERSIONS

commit: None

pandas: 0.24.2
pytest: None
pip: 19.1.1
setuptools: 41.0.1
Cython: 0.29.10
numpy: 1.16.4
scipy: 1.2.1
pyarrow: None
xarray: None
IPython: 7.5.0
sphinx: None
patsy: 0.5.1
dateutil: 2.8.0
pytz: 2019.1
blosc: None
bottleneck: None
tables: None
numexpr: None
feather: None
matplotlib: 3.1.0
openpyxl: None
xlrd: None
xlwt: None
xlsxwriter: None
lxml.etree: None
bs4: None
html5lib: None
sqlalchemy: None
pymysql: None
psycopg2: 2.7.5 (dt dec pq3 ext lo64)
jinja2: 2.10.1
s3fs: None
fastparquet: None
pandas_gbq: None
pandas_datareader: None
gcsfs: None

[TomAugspurger]

I can't reproduce locally. Is the prophet import necessary?

[TomAugspurger]

@Dharni0607 pandas will import Matplotlib internally. And plt.plot() may / will be different from Series.plot.

[Dharni0607]

sorry for creating confusion.

[TomAugspurger]

No worries!

…

On Jun 26, 2019, at 07:55, Dharni0607 ***@***.***> wrote: sorry for creating confusion. — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or mute the thread.

[Dharni0607]

I have taken the code from this documentation:
https://pandas.pydata.org/pandas-docs/stable/getting_started/10min.html#plotting

I was getting similar error without importing Prophet as well.

---

[TomAugspurger]

OK, if someone who can reproduce it locally can debug it we can figure out what's going wrong.

…

On Wed, Jun 26, 2019 at 10:13 AM Dharni0607 ***@***.***> wrote: I have taken the code from this documentation: https://pandas.pydata.org/pandas-docs/stable/getting_started/10min.html#plotting I was getting similar error without importing Prophet as well. [image: ts] <https://user-images.githubusercontent.com/30770547/60191543-1a731280-9852-11e9-8f4a-63aaa3372067.png> ------------------------------ — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#27036?email_source=notifications&email_token=AAKAOITFEZSZU2ZE2YDWSBLP4OBSDA5CNFSM4H3IKCO2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODYT3RYY#issuecomment-505919715>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAKAOIR7WEMGHUA5JRSNJMTP4OBSDANCNFSM4H3IKCOQ> .

[Dharni0607]

I think this issue can be related to this:
matplotlib/matplotlib#14359

---

[TomAugspurger]

Ah, so the reproducer is

import pandas as pd
import numpy as np

from pandas.plotting import deregister_matplotlib_converters

deregister_matplotlib_converters()


ts = pd.Series(np.random.randn(1000), index=pd.date_range('1/1/2000', periods=1000))
ts.plot()

~/sandbox/pandas/pandas/plotting/_core.py in __call__(self, kind, ax, figsize, use_index, title, grid, legend, style, logx, logy, loglog, xticks, yticks, xlim, ylim, rot, fontsize, colormap, table, yerr, xerr, label, secondary_y, **kwds)
    808                            colormap=colormap, table=table, yerr=yerr,
    809                            xerr=xerr, label=label, secondary_y=secondary_y,
--> 810                            **kwds)
    811     __call__.__doc__ = plot_series.__doc__
    812

~/sandbox/pandas/pandas/plotting/_core.py in plot_series(data, kind, ax, figsize, use_index, title, grid, legend, style, logx, logy, loglog, xticks, yticks, xlim, ylim, rot, fontsize, colormap, table, yerr, xerr, label, secondary_y, **kwds)
    766                  yerr=yerr, xerr=xerr,
    767                  label=label, secondary_y=secondary_y,
--> 768                  **kwds)
    769
    770

~/sandbox/pandas/pandas/plotting/_core.py in _plot(data, x, y, subplots, ax, kind, **kwds)
    714         plot_obj = klass(data, subplots=subplots, ax=ax, kind=kind, **kwds)
    715
--> 716     plot_obj.generate()
    717     plot_obj.draw()
    718     return plot_obj.result

~/sandbox/pandas/pandas/plotting/_matplotlib/core.py in generate(self)
    214         self._compute_plot_data()
    215         self._setup_subplots()
--> 216         self._make_plot()
    217         self._add_table()
    218         self._make_legend()

~/sandbox/pandas/pandas/plotting/_matplotlib/core.py in _make_plot(self)
    987                              stacking_id=stacking_id,
    988                              is_errorbar=is_errorbar,
--> 989                              **kwds)
    990             self._add_legend_handle(newlines[0], label, index=i)
    991

~/sandbox/pandas/pandas/plotting/_matplotlib/core.py in _ts_plot(cls, ax, x, data, style, **kwds)
   1025         ax._plot_data.append((data, cls._kind, kwds))
   1026
-> 1027         lines = cls._plot(ax, data.index, data.values, style=style, **kwds)
   1028         # set date formatter, locators and rescale limits
   1029         format_dateaxis(ax, ax.freq, data.index)

~/sandbox/pandas/pandas/plotting/_matplotlib/core.py in _plot(cls, ax, x, y, style, column_num, stacking_id, **kwds)
   1002             cls._initialize_stacker(ax, stacking_id, len(y))
   1003         y_values = cls._get_stacked_values(ax, stacking_id, y, kwds['label'])
-> 1004         lines = MPLPlot._plot(ax, x, y_values, style=style, **kwds)
   1005         cls._update_stacker(ax, stacking_id, y)
   1006         return lines

~/sandbox/pandas/pandas/plotting/_matplotlib/core.py in _plot(cls, ax, x, y, style, is_errorbar, **kwds)
    581             else:
    582                 args = (x, y)
--> 583             return ax.plot(*args, **kwds)
    584
    585     def _get_index_name(self):

~/Envs/pandas-dev/lib/python3.7/site-packages/matplotlib/axes/_axes.py in plot(self, scalex, scaley, data, *args, **kwargs)
   1666         lines = [*self._get_lines(*args, data=data, **kwargs)]
   1667         for line in lines:
-> 1668             self.add_line(line)
   1669         self.autoscale_view(scalex=scalex, scaley=scaley)
   1670         return lines

~/Envs/pandas-dev/lib/python3.7/site-packages/matplotlib/axes/_base.py in add_line(self, line)
   1898             line.set_clip_path(self.patch)
   1899
-> 1900         self._update_line_limits(line)
   1901         if not line.get_label():
   1902             line.set_label('_line%d' % len(self.lines))

~/Envs/pandas-dev/lib/python3.7/site-packages/matplotlib/axes/_base.py in _update_line_limits(self, line)
   1920         Figures out the data limit of the given line, updating self.dataLim.
   1921         """
-> 1922         path = line.get_path()
   1923         if path.vertices.size == 0:
   1924             return

~/Envs/pandas-dev/lib/python3.7/site-packages/matplotlib/lines.py in get_path(self)
   1025         """
   1026         if self._invalidy or self._invalidx:
-> 1027             self.recache()
   1028         return self._path
   1029

~/Envs/pandas-dev/lib/python3.7/site-packages/matplotlib/lines.py in recache(self, always)
    668         if always or self._invalidx:
    669             xconv = self.convert_xunits(self._xorig)
--> 670             x = _to_unmasked_float_array(xconv).ravel()
    671         else:
    672             x = self._x

~/Envs/pandas-dev/lib/python3.7/site-packages/matplotlib/cbook/__init__.py in _to_unmasked_float_array(x)
   1388         return np.ma.asarray(x, float).filled(np.nan)
   1389     else:
-> 1390         return np.asarray(x, float)
   1391
   1392

~/Envs/pandas-dev/lib/python3.7/site-packages/numpy/core/numeric.py in asarray(a, dtype, order)
    536
    537     """
--> 538     return array(a, dtype, copy=False, order=order)
    539
    540

TypeError: float() argument must be a string or a number, not 'Period'

We should figure something out.

[Dharni0607]

I think, plot() fails to work with formatters and converters after deregistering.

This is the source:
https://github.com/pandas-dev/pandas/blob/v0.24.2/pandas/plotting/_converter.py#L88-L114

def deregister():

    Remove pandas' formatters and converters
    Removes the custom converters added by :func:`register`. This
    attempts to set the state of the registry back to the state before
    pandas registered its own units. Converters for pandas' own types like
    Timestamp and Period are removed completely. Converters for types
    pandas overwrites, like ``datetime.datetime``, are restored to their
    original value. 

Without date_range(), it is giving a decent plot

import pandas as pd
import numpy as np
from pandas.plotting import deregister_matplotlib_converters
deregister_matplotlib_converters()
ts = pd.Series(np.random.randn(1000), index = range(1000))   
ts.plot()

[nthanhtin]

A temporary fix if you still want to have date_range():

from pandas.plotting import register_matplotlib_converters       

register_matplotlib_converters()       

ts.plot() 

[amerberg]

I believe this bug is caused by the following two factors:

1. register_matplotlib_converters is not idempotent. Calling the function once stores the existing converters in a cache. Calling it a second time overwrites that cache with the newly-registered converters.

2. Importing anything from pandas.plotting._core forces a call to register_matplotlib_converters.

The first call to register_matplotlib_converters caches the default matplotlib converters. Then the second call overwrites them with the pandas converters. And then when we go to deregister, there's a check that prevents pandas converters from being replaced. So we end up with no converter at all.

I propose to avoid caching the pandas converters in the register function. This will eliminate the need for the check in the deregister function. I'll start working on a pull request.

[amerberg]

I believe this is a duplicate of #27479.

[TomAugspurger]

I think it's only tangentially related to #27479

I think the best course forward is for prophet disable the converts before plotting, and reenable them (if they were previously enabled) after.