通过替换不同列中的nans来合并数据框内的行

Question

I have a df:

df = pd.DataFrame([[1, np.nan, "filled", 3], [1, "filled", np.nan, 3], [1, "filled", np.nan, 4]], columns = ["a", "b", "c", "d"])
    a   b   c   d
0   1   NaN filled  3
1   1   filled  NaN 3
2   1   filled  NaN 4

And my end result should be:

df = pd.DataFrame([[1, "filled", "filled", 3], [1, "filled", np.nan, 4]], columns = ["a", "b", "c", "d"])
    a   b   c   d
0   1   filled  filled  3
1   1   filled  NaN 4

So I want to merge the rows that are identical in all respects other than the column b and c. The issue is that not always there will be a another row identical except for columns b and c.

Can't think how to use df.groupby(["a", "d"]).apply() to get what I want.

Answer 1

You can check with groupby + first , it will select the first not NaN value as output

df.groupby(['a','d'],as_index=False).first()
Out[897]: 
   a  d       b       c
0  1  3  filled  filled
1  1  4  filled     NaN