ENH: df.add fill_value NotImplementedError

[neverforgit]

It appears that fill_value option for the df.add() method is not actually implemented in the source code. This was brought up in a StackOverflow question.

piRSquared is the user that identified it.

The following are all copied from that SO post.

Code Sample, a copy-pastable example if possible

import pandas as pd
import numpy as np

np.random.seed([3,1415])
df = pd.DataFrame(np.random.choice((1, 2, np.nan), (5, 5)))
s = pd.Series(range(5))
df.add(s, fill_value=0)
>>>
  ---------------------------------------------------------------------------
  NotImplementedError                       Traceback (most recent call last)
  <ipython-input-18-ce2809aeda79> in <module>()
  ----> 1 df.add(s, fill_value=0)

  /Users/daddy30000/anaconda/lib/python2.7/site-packages/pandas/core/ops.pyc in f(self, other, axis, level, fill_value)
     1060             return self._combine_frame(other, na_op, fill_value, level)
     1061         elif isinstance(other, ABCSeries):
  -> 1062             return self._combine_series(other, na_op, fill_value, axis, level)
     1063         elif isinstance(other, (list, tuple)):
     1064             if axis is not None and self._get_axis_name(axis) == 'index':

  /Users/daddy30000/anaconda/lib/python2.7/site-packages/pandas/core/frame.pyc in _combine_series(self, other, func, fill_value, axis, level)
     3511             else:
     3512                 return self._combine_match_columns(other, func, level=level,
  -> 3513                                                    fill_value=fill_value)
     3514         return self._combine_series_infer(other, func, level=level,
     3515                                           fill_value=fill_value)

  /Users/daddy30000/anaconda/lib/python2.7/site-packages/pandas/core/frame.pyc in _combine_match_columns(self, other, func, level, fill_value)
     3542         if fill_value is not None:
     3543             raise NotImplementedError("fill_value %r not supported" %
  -> 3544                                       fill_value)
     3545
     3546         new_data = left._data.eval(func=func, other=right,

  NotImplementedError: fill_value 0 not supported

Expected Output

   0    1     2     3    4  
0  1.0  1.0  2.0  3.0  4.0  
1  2.0  3.0  2.0  4.0  4.0  
2  1.0  1.0  3.0  4.0  4.0  
3  1.0  1.0  2.0  4.0  6.0  
4  1.0  3.0  4.0  3.0  5.0

output of pd.show_versions()

INSTALLED VERSIONS
------------------
commit: None
python: 2.7.11.final.0
python-bits: 64
OS: Darwin
OS-release: 15.5.0
machine: x86_64
processor: i386
byteorder: little
LC_ALL: None
LANG: en_US.UTF-8

pandas: 0.18.1
nose: 1.3.7
pip: 8.1.1
setuptools: 21.2.1
Cython: 0.22.1
numpy: 1.11.0
scipy: 0.15.1
statsmodels: 0.6.1
xarray: None
IPython: 3.2.0
sphinx: 1.3.1
patsy: 0.3.0
dateutil: 2.5.3
pytz: 2016.4
blosc: None
bottleneck: 1.0.0
tables: 3.2.0
numexpr: 2.4.3
matplotlib: 1.4.3
openpyxl: 1.8.5
xlrd: 0.9.3
xlwt: 1.0.0
xlsxwriter: 0.7.3
lxml: 3.4.4
bs4: 4.3.2
html5lib: None
httplib2: None
apiclient: None
sqlalchemy: 1.0.5
pymysql: None
psycopg2: 2.5.4 (dt dec pq3 ext)
jinja2: 2.7.3
boto: 2.38.0
pandas_datareader: None

[jreback]

simply df.fillna(0).add(s).

I'll mark it, but these things are generally would need outside patches.

[MTKnife]

simply df.fillna(0).add(s).

That works for some use cases, but it doesn't work when you want to add two objects with different indices. E.g.:

foo = DataFrame([[1, 2, 3], [4, 5, 6]])
bar = Series([7, 8, 9, 10])

If I want to add these two together, I'd like to be able to assume that missing columns in "foo" are 0 values (it's the kind of thing that can happen easily when you're adding new data to old data--in my case, a vocabulary that expands over time), like this:

foo.add(bar, axis = 0, fill_value = 0)

Yes, I can still do this by reindexing "foo" before calling "add", but that involves more work than a simple "fillna") call. Not only is it a pain, but it prevents me from taking advantage of polymorphism: for example, right now I'm working on an application where the argument of an "add" can be either a DataFrame or a Series, and the app worked fine while the argument was a DataFrame, leaving me with a bug that I discovered only when I accidentally supplied a set of arguments that created a Series.

[jonathanrhughes]

take

[mroeschke]

Superseded by #32608 so closing