Given the following:
df = pd.DataFrame({\'col1\' : [\"a\",\"b\"], \'col2\' : [\"ab\",np.nan], \'col3\' : [\"w\",\"e\"]})
I would l
In [1556]: df.apply(lambda x: '*'.join(x.dropna().astype(str).values), axis=1) Out[1556]: 0 a*ab*w 1 b*e 2 3*4*� 3 ñ*ü*á dtype: object