Fill NULL based on conditions python-CodePudding

I have a dataset, and needs to create a new column to deal with the NULL based on the conditions. Here is the sample data:

id1 id2 col1 col2 
1   1   2     3 
2   2   2     NULL
3   3   NULL  3
4   4   NULL  3
5   5   10    11

The expect output is

id1 id2 col1 col2  combine
1   1   2     3     2,3
2   2   2     NULL  2,3
3   3   NULL  3     2,3
4   4   NULL  3     2,3
5   5   10    11    10,11

for example,If any col1 and col2 has 2 or 3, the new column need to contain both values.Not sure how to capture if col1 or col2 has NULLs.

CodePudding user response：

You can use ffill to fill NaNs convert to int, then str, then join on axis:

df['combine'] = df[['col1','col2']].ffill().astype(int).astype(str).apply(','.join, axis=1)

Output:

   id1  id2  col1  col2 combine
0    1    1   2.0   3.0     2,3
1    2    2   2.0   NaN     2,3
2    3    3   NaN   3.0     2,3
3    4    4  10.0  11.0   10,11