I want to replace a NaN in a dataframe column by a dictionary like this: {"value":["100"]}
df[column].apply(type).value_counts()
output:
<class 'dict'> 11565
<class 'float'> 43
df[column].isna().sum()
output => 43
How can I do this?
Use lambda function for replace by dictionary:
df = pd.DataFrame({'column':[np.nan, {'a':[4,5]}]})
d = {"value":["100"]}
df['column'] = df['column'].apply(lambda x: d if pd.isna(x) else x)
print (df)
column
0 {'value': ['100']}
1 {'a': [4, 5]}
Or list comprehension:
df['column'] = [d if pd.isna(x) else x for x in df['column']]
