In dataframe, move data from column2 to column1, if column1 is empty-CodePudding

move data from column2 to column1, if column1 is empty

In id 2 and 4 Phone1 is empty so transfer data from phone2 to phone1 df

id  Phone1      Phone2
0   9073370404
1   9072765522  1234567890
2               9075611434
3   9073370404
4               9072765522
5   9075611434

id  Phone1      Phone2
0   9073370404
1   9072765522  1234567890
2   9075611434  
3   9073370404
4   9072765522  
5   9075611434

CodePudding user response：

Firstly, create the dataframe

import pandas as pd
import numpy as np
df = pd.DataFrame({
"Phone1":[9073370404, 9072765522, None, 9073370404, None, 9075611434],
"Phone2": [None, 1234567890, 9075611434, None, 9072765522, None]
})
print(df)

Secondly, transfer data from phone2 to phone1

df["Phone1"] = df.apply(lambda row: row["Phone2"] if ((row["Phone1"] is None) or np.isnan(row["Phone1"])) else row["Phone1"], axis=1)
print(df)

Hope help you

CodePudding user response：

@Lowin Li

is not set the df["Phone2"], and import numpy

you can use apply like @Lowin Li

import pandas as pd
dict = [
{"id":"0","Phone1":"9073370404",},
{"id":"1","Phone1":"9072765522","Phone2":"1234567890"},
{"id":"2","Phone2":"9075611434"},
{"id":"3","Phone1":"9073370404",},
{"id":"4","Phone2":"9072765522"},
{"id":"5","Phone1":"9075611434",}
]

df = pd.DataFrame.from_dict(dict)
def my_conbin(a, b):
    if pd.isna(a):
        return b
    return a
def my_delete(a, b):
    if a == b:
        return None
    return b
df['Phone1'] = df.apply(lambda row: my_conbin(row['Phone1'], row['Phone2']), axis=1)
df['Phone2'] = df.apply(lambda row: my_delete(row['Phone1'], row['Phone2']), axis=1)
print(df)

and df.iterrows() is more simple:

import pandas as pd
dict = [
{"id":"0","Phone1":"9073370404",},
{"id":"1","Phone1":"9072765522","Phone2":"1234567890"},
{"id":"2","Phone2":"9075611434"},
{"id":"3","Phone1":"9073370404",},
{"id":"4","Phone2":"9072765522"},
{"id":"5","Phone1":"9075611434",}
]

df = pd.DataFrame.from_dict(dict)

for index,row in df.iterrows():
    if pd.isna(row.Phone1):
        df.loc[index, 'Phone1']= row.Phone2
        df.loc[index, 'Phone2'] = None
print(df)

CodePudding user response：

def function1(s:pd.Series):
    if pd.isna(s.Phone1):
        s.Phone1,s.Phone2=s.Phone2,s.Phone1
    return s

df1.apply(function1,axis=1)