How to choose row from other column in dataframe if row from the default column is NaN?

user_18457560 0
Viewed 42

My dataframe consists of 3 columns. The thirth column is based on the first two columns. The default column is column 2. But if column 2 is NaN, then I want column 3 to be filled with column 1. I added the third line to conditions, but it does not seem to work.

This is the DataFrame:

df = pd.DataFrame(np.array([[np.nan, 1717], [1749, 1750], [1704, np.nan]]),
                   columns=['a', 'b'])

This is my code:

import numpy as np
import pandas as pd
conditions = [
    (df["b"] <= df["a"]), 
    df["b"] > df["a"],
    df["b"] == df["b"].isna()]

choices = [df["b"], df["a"], df["a"]]

df['c'] = np.select(conditions, choices, default=df["b"])
print(df)

This is my output:

           a            b      c
0        NaN         1749.0  1749.0
1        1717.0      1750.0  1717.0
2        1704.0      NaN     NaN

But I want c to be filled if a or b is filled. So this is the output I want:

           a            b      c
0        NaN         1749.0  1749.0
1        1717.0      1750.0  1717.0
2        1704.0      NaN     1704.0
pandas dataframe nan numpy python
3 Answers

You just need to make a small change to your third condition. df["b"].isna() already returns True or False, so df["b"] == df["b"].isna() is actually checking to see if df["b"] evaluates to the same boolean (it doesn't).

Just remove the first half of the third condition.

import numpy as np
import pandas as pd
conditions = [
    (df["b"] <= df["a"]), 
    df["b"] > df["a"],
    df["b"].isna()]

choices = [df["b"], df["a"], df["a"]]

df['c'] = np.select(conditions, choices, default=df["b"])
print(df)
user_131871040

This seems to work:

df = pd.DataFrame(np.array([[np.nan, 1717], [1749, 1750], [1704, np.nan]]),
               columns=['a', 'b'])

df['c'] = df.a

for i in range(len(df)):
    if df.a.iloc[i] == np.nan:
        df.c.iloc[i] = df.b.iloc[i]
user_24291870

This solution gives the output that you want:

import pandas as pd
import numpy as np

df = pd.DataFrame(np.array([[np.nan, 1717], [1749, 1750], [1704, np.nan]]),
                   columns=['a', 'b'])

def fill_row(row):
    if pd.isnull(row['a']): 
        return row['b']
    else: 
        return row['a']

df['c'] = df.apply(lambda row : fill_row(row), axis=1) 
print(df)

The output:

        a       b       c
0     NaN  1717.0  1717.0
1  1749.0  1750.0  1749.0
2  1704.0     NaN  1704.0
user_113044900
Related