how to fill an empty dataframe column in pandas python

Question

I m using Pandas with Python 3.6. My script loads an excel file which contains multiple worksheets. In some sheets, rows have either numeric values, or string values on two columns. After running the script, the numeric values are splitted on two columns, but I cannot duplicate on the second column the string value of the first column.

For the numeric values, I am using :

df=df[['ID_Test']].join(df[pd_column].str.split(':',expand=True)).rename(columns={0: pd_column, 1: ''})

The second column for the string values remains blank (None), and must be updated with the same value than the first column

If I use : df[''] = df[pd_column] , the second column [''] is entirely updated with the values of the first one (overwriting numeric values), and I did not find any solution specific to my concern.

Data Input:

ID_Test_1   Test_1
Indicator_1 AAAAAAA
Indicator_2 2.745 : 2.03
Indicator_3 BBBBBBBB
Indicator_4 -5.013 : -5.013
Indicator_5 CCCCCCCC

Actual Output : (Wrong)

ID_Test_1   Test_1  
Indicator_1 AAAAAAA      None
Indicator_2 2.745        2.03
Indicator_3 BBBBBBBB     None
Indicator_4 -5.013       -5.013
Indicator_5 CCCCCCCCC    None

Desired Output :

ID_Test_1   Test_1  
Indicator_1 AAAAAAA      AAAAAAA
Indicator_2 2.745        2.03
Indicator_3 BBBBBBBB     BBBBBBBB
Indicator_4 -5.013       -5.013
Indicator_5 CCCCCCCCC    CCCCCCCCC

The second column must not have a label

Answer 1

Base on your sample data and code. I just added ffill(1) to the split before join

pd_column = 'Test_1'
(df[['ID_Test_1']].join(df[pd_column].str.split('\s+:\s+',expand=True).ffill(1))
                                    .rename(columns={0: pd_column, 1: ''}))


Out[29]:
     ID_Test_1    Test_1
0  Indicator_1   AAAAAAA   AAAAAAA
1  Indicator_2     2.745      2.03
2  Indicator_3  BBBBBBBB  BBBBBBBB
3  Indicator_4    -5.013    -5.013
4  Indicator_5  CCCCCCCC  CCCCCCCC