Fill NaN values based on another column

Question

I have this pd.DataFrame:

test4 = pd.DataFrame({'condition': ['good', np.nan, np.nan, 'excellent', 'good', np.nan], 
                      'odometer': [np.nan, 35000, 20000, 100000, 500000, 50]})

Output:

condition   odometer
0   good    NaN
1   NaN 35000.0
2   NaN 20000.0
3   excellent   100000.0
4   good    500000.0
5   NaN 50.0

And I can't fill the NaN values in the column "condition" using the values in the column "odometer". The conditions would be:

odometer <=30000, then condition = 'good'
odometer > 30000 & odometer <=150000, then value = 'excellent'
odometer > 150000 & odometer <=10000000, then value = 'good'

And it should look like this in the end:

condition   odometer
0   good    NaN
1   excellent   35000.0
2   good    20000.0
3   excellent   100000.0
4   good    500000.0
5   good    50.0

I have tried different stuff but none worked. For example:

def f(change):
    if change['condition'] == np.nan:
        condition = change['odometer']
        value = change['condition']
        if   condition <=30000 value = 'good'
        elif condition > 30000 & condition <=150000 value = 'excellent'
        elif condition > 150000 & condition <=10000000 value = 'good'
        return value
    return change['condition']

test4['condition'] = test4.apply(f)

What am I doing wrong? Is there a way to make it work?

Answer 1

I found the answer:

m1 = (test4['odometer'] > 30000) & (test4['odometer'] <= 150000)
m2 = (test4['odometer'] <= 30000) | (test4['odometer'] > 150000)

test4.loc[m1,'condition'] = test4.loc[m1,'condition'].fillna('excellent')
test4.loc[m2,'condition'] = test4.loc[m2,'condition'].fillna("good")