intersection de dataframes basée sur la colonne

s1 = pd.merge(dfA, dfB, how='inner', on=['S', 'T'])
s1.dropna(inplace=True)
Real Raccoon