df.dropna(subset=['column1','column2'], inplace=True)
df = df[df['EPS'].notna()] #instead of dropping, take another DF with non-na values
df.dropna(subset = ['column1', 'column2'], inplace = True)