Superimpose on a data frame-CodePudding

A is my reference data-frame. It will always have only one row.

A = pd.DataFrame({'a': [5], 'b': [3], 'c': ['C'], 'd' : ['D'], 'e' : ['E'], 'f' : [True], 'g' : [7]})

I want to use above data frame to replace NA and append missing columns in the below data frame. The index of A is ignored.

x = pd.DataFrame({'a': [1,2,3,4], 'b': [0,1,np.nan, 1], 'e': ['Ex1', np.nan, 'Ex2', np.nan]})

The expected result is shown below.

Language - Python, Library pandas

CodePudding user response：

You can reindex using the columns of A and then .fillna using the first (and only) row (as a Series) so that the filling aligns on the column labels.

Let pandas 'infer' the downcasting so that the rows that can be cast to int64 don't remain as float64, if that's important.

x = x.reindex(A.columns, axis=1).fillna(A.iloc[0], downcast='infer')

   a  b  c  d    e     f  g
0  1  0  C  D  Ex1  True  7
1  2  1  C  D    E  True  7
2  3  3  C  D  Ex2  True  7
3  4  1  C  D    E  True  7