from sklearn import preprocessing le = preprocessing.LabelEncoder() df['label'] = le.fit_transform(df['label'])