CODEWITHSUNDEEP
pythonpandas
sachinruk
sachinruk

Reputation: 9869

Set list in subset of pandas dataframe

This is similar to this question, however differs in that I am concerned about a subset of the dataframe.

Suppose I have the following dataframe:

import pandas as pd
import numpy as np

np.random.seed(42)
df = pd.DataFrame(np.random.randn(5), columns=['A'])

and I wish to have a column 'B' that has the value ["neg"] for the negative values of A. However, when I do the following as was suggested in the linked question, I lose the list nature of what I was expecting.

idx = df.A < 0  
df.loc[idx, "B"] = [["neg"]]*idx.sum() 
>>>
Out[17]: 
          A    B
0  0.496714  NaN
1 -0.138264  neg
2  0.647689  NaN
3  1.523030  NaN
4 -0.234153  neg

What am I doing wrong here? The only thing I can do in the meantime to fix this is to do: df.loc[idx, 'B'] = df.loc[idx, 'B'].map(lambda x: [x]).

Upvotes: 1

Views: 577

Answers (1)

mechanical_meat
mechanical_meat

Reputation: 169434

You need to provide a pd.Series, and you must supply the length of the entire DataFrame to associate ["neg"] to any of the negative value(s):

df.loc[idx, "B"] = pd.Series([["neg"]]*len(df))

Result:

          A      B
0  0.496714    nan
1 -0.138264  [neg]
2  0.647689    nan
3  1.523030    nan
4 -0.234153  [neg]

Upvotes: 1

Related Questions