Get column names from ColumnTransformer with embedded pipelines (SKLearn)

ColumnTransform_pipeline_field_extraction.py

## Split into holdout for purposes of imputation and encoding
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = .2, random_state=2019)
class_labels = df['segment_label'].cat.categories

## Preprocessing pipeline

# Define transforms on numeric types
numeric_features = X.select_dtypes(np.number).columns
numeric_transformer = Pipeline(steps=[
    ('imputer', SimpleImputer(strategy='median')),
    ('scaler', StandardScaler())])

# Define transforms on categorical types
categorical_features = X.select_dtypes(['object', 'bool', 'category']).columns
categorical_transformer = Pipeline(steps=[
    ('imputer', SimpleImputer(strategy='constant', fill_value='missing')),
    ('onehot', OneHotEncoder(handle_unknown='ignore'))])

# Construct ColumnTransformer object
preprocessor = ColumnTransformer(
    transformers=[
        ('num', numeric_transformer, numeric_features),
        ('cat', categorical_transformer, categorical_features)
    ],
    remainder='drop',
    verbose=True)

# Visual inspection of transformed dataframe before CV
preprocessor.fit_transform(X_train, y_train)
enc_cat_features = preprocessor.named_transformers_['cat']['onehot'].get_feature_names()
labels = np.concatenate([numeric_features, enc_cat_features])
transformed_df_columns = pd.DataFrame(preprocessor.transform(X_train).toarray(), columns=labels).columns
pp.pprint(transformed_df_columns)