Update column selector

DataCanvasIO · Feb 15, 2024 · f529541 · f529541
1 parent 903da91
commit f529541
Show file tree

Hide file tree

Showing 2 changed files with 3 additions and 3 deletions.
diff --git a/hypernets/tabular/cuml_ex/_data_cleaner.py b/hypernets/tabular/cuml_ex/_data_cleaner.py
@@ -30,7 +30,7 @@ def as_local(self):
                              reduce_mem_usage=self.reduce_mem_usage,
                              int_convert_to=self.int_convert_to)
         copy_attrs_as_local(self, target, 'df_meta_', 'columns_', 'dropped_constant_columns_',
-                   'dropped_idness_columns_', 'dropped_duplicated_columns_')
+                            'dropped_idness_columns_', 'dropped_duplicated_columns_')
 
         return target
 
@@ -52,7 +52,7 @@ def _get_duplicated_columns(df):
 
     @staticmethod
     def replace_nan_chars(X: cudf.DataFrame, nan_chars):
-        cat_cols = X.select_dtypes(['object', ])
+        cat_cols = X.select_dtypes(['object', 'string', ])
         if cat_cols.shape[1] > 0:
             cat_cols = cat_cols.replace(nan_chars, cupy.nan)
             X[cat_cols.columns] = cat_cols

diff --git a/hypernets/tabular/dask_ex/_transformers.py b/hypernets/tabular/dask_ex/_transformers.py
@@ -292,7 +292,7 @@ def fit(self, X, y=None):
         self.dtypes_ = {c: X[c].dtype for c in X.columns}
 
         if self.columns is None:
-            columns = X.select_dtypes(include=["category", 'object', 'bool']).columns.to_list()
+            columns = X.select_dtypes(include=['category', 'object', 'string', 'bool']).columns.to_list()
         else:
             columns = self.columns