[Python] How to select and transform cols with sklearn pipeline and preserver original names?

Here, the purpose is to select a few cols from a dataframe. Then apply log transformation some of the selected cols.

log_transformer = FunctionTransformer(lambda x: np.log2(x + 1), validate=False, feature_names_out="one-to-one")

preprocessor = ColumnTransformer([
("selector", "passthrough", selected_cols),
("log_transform", log_transformer, log_cols),
], remainder="drop", verbose_feature_names_out=False)

pipeline = Pipeline([
("preprocessor", preprocessor)
]).fit(train_df)


Here, the software ask me to set verbose_feature_names_out=True. Is there a way to prevent "selector" and "log_transform" prefixes to be added to the col names?

Continue reading...