pandas.DataFrame.plot error for time series data

[alexgawrilow]

Code Sample

import pandas as pd

idx = pd.date_range(start="2018", freq=pd.DateOffset(days=1), periods=2)
ts = pd.DataFrame({"A": [1, 2]}, idx)
ts.plot()

Problem description

I created the index of a time series using pd.date_range and provided the frequency as a DateOffset object. This index was passed to the DataFrame constructor to create a time series. When calling the plot method on this time series, pandas is not able to parse the frequency correctly:

233     if isinstance(freq, DateOffset):
--> 234         freq = freq.rule_code
    235     else:
    236         freq = get_base_alias(freq)

~/venv/lib/python3.6/site-packages/pandas/tseries/offsets.py in rule_code(self)
    372     @property
    373     def rule_code(self):
--> 374         return self._prefix
    375 
    376     @cache_readonly

~venv/lib/python3.6/site-packages/pandas/tseries/offsets.py in _prefix(self)
    368     @property
    369     def _prefix(self):
--> 370         raise NotImplementedError('Prefix not defined')
    371 
    372     @property

Expected Output

When passing the frequency to pd.date_range as a string, everything works fine, however both indices are equal.

idx2 = pd.date_range(start="2018", freq="D", periods=2)
ts = pd.DataFrame({"A": [1, 2]}, idx2)
ts.plot()
idx.equals(idx2)

Output: True

Output of pd.show_versions()

INSTALLED VERSIONS

commit: None
python: 3.6.7.final.0
python-bits: 64
OS: Linux
OS-release: 4.18.0-10-generic
machine: x86_64
processor: x86_64
byteorder: little
LC_ALL: None
LANG: en_US.UTF-8
LOCALE: en_US.UTF-8

pandas: 0.24.2
pytest: 4.3.1
pip: 19.1.1
setuptools: 41.0.1
Cython: 0.29.6
numpy: 1.16.2
scipy: None
pyarrow: None
xarray: None
IPython: 7.3.0
sphinx: None
patsy: None
dateutil: 2.8.0
pytz: 2018.9
blosc: None
bottleneck: None
tables: None
numexpr: None
feather: None
matplotlib: 2.2.2
openpyxl: None
xlrd: None
xlwt: None
xlsxwriter: None
lxml.etree: None
bs4: 4.7.1
html5lib: None
sqlalchemy: None
pymysql: None
psycopg2: None
jinja2: 2.10.1
s3fs: None
fastparquet: None
pandas_gbq: None
pandas_datareader: None
gcsfs: None

[TomAugspurger]

Can you look to see what the plotting code expects about the freq? I think in the spot it raises, it's checking to see if the freq is fixed or variable. And if it's variable it bails out.

I think for arbitrary DateOffsets, which may not have a fixed frequency, we can't expect to use pandas formatters. What do you think?

[kmw8551]

I'll take a look

[MarcoGorelli]

Full traceback:

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/home/marco/pandas-dev/pandas/plotting/_core.py", line 907, in __call__
    return plot_backend.plot(data, kind=kind, **kwargs)
  File "/home/marco/pandas-dev/pandas/plotting/_matplotlib/__init__.py", line 61, in plot
    plot_obj.generate()
  File "/home/marco/pandas-dev/pandas/plotting/_matplotlib/core.py", line 264, in generate
    self._make_plot()
  File "/home/marco/pandas-dev/pandas/plotting/_matplotlib/core.py", line 1057, in _make_plot
    if self._is_ts_plot():
  File "/home/marco/pandas-dev/pandas/plotting/_matplotlib/core.py", line 1049, in _is_ts_plot
    return not self.x_compat and self.use_index and self._use_dynamic_x()
  File "/home/marco/pandas-dev/pandas/plotting/_matplotlib/core.py", line 1054, in _use_dynamic_x
    return _use_dynamic_x(self._get_ax(0), self.data)
  File "/home/marco/pandas-dev/pandas/plotting/_matplotlib/timeseries.py", line 205, in _use_dynamic_x
    freq = get_period_alias(freq)
  File "/home/marco/pandas-dev/pandas/plotting/_matplotlib/timeseries.py", line 167, in get_period_alias
    freq = freq.rule_code
  File "pandas/_libs/tslibs/offsets.pyx", line 532, in pandas._libs.tslibs.offsets.BaseOffset.rule_code.__get__
    return self._prefix
  File "pandas/_libs/tslibs/offsets.pyx", line 528, in pandas._libs.tslibs.offsets.BaseOffset._prefix.__get__
    raise NotImplementedError("Prefix not defined")
NotImplementedError: Prefix not defined

Is this just the case of modifying _use_dynamic_x so that it would return False here, as it would in the case

import pandas as pd

idx = pd.date_range(start="2018", freq=pd.offsets.WeekOfMonth(), periods=2)
ts = pd.DataFrame({"A": [1, 2]}, idx)
ts.plot()

?