# Option 1 df1 = df.replace(np.nan, '', regex=True) # All data frame # Option 2 df[['column1','column2']] = df[['column1','column2']].fillna('') # Specific columns