使用非NaN的填充值初始化Pandas DataFrame

Question

Suppose I initialize an 'empty' DataFrame as follows:

import pandas as pd
import numpy as np
df = pd.DataFrame(index=list('AB'), columns=list('CD'))

The resulting df has the form

     C    D
A  NaN  NaN
B  NaN  NaN

Is there a pythonic way to replace the NaN s with some other value, say, -np.inf ? Of course, one way to do it is simply to specify it as data:

df = pd.DataFrame(data=np.ones((2,2))*(-np.inf), index=list('AB'), columns=list('CD'))

Perhaps there a more succinct way?

Answer 1

pass scalar value as data param, this will set all elements to the same value:

In [181]:
df = pd.DataFrame(index=list('AB'), columns=list('CD'), data=-np.inf)
df

Out[181]:
     C    D
A -inf -inf
B -inf -inf

The docs show that data param can accept constants (scalar values) as well as array-like structures and dicts.