Merge pull request #16856 from rapidsai/branch-24.10

Forward-merge branch-24.10 into branch-24.12

python/cudf/cudf/_lib/concat.pyx

@@ -23,9 +23,9 @@ def concat_columns(object columns):
 def concat_tables(object tables, bool ignore_index=False):
     plc_tables = []
     for table in tables:
-        cols = table._data.columns
+        cols = table._columns
         if not ignore_index:
-            cols = table._index._data.columns + cols
+            cols = table._index._columns + cols
         plc_tables.append(pylibcudf.Table([c.to_pylibcudf(mode="read") for c in cols]))
 
     return data_from_pylibcudf_table(

python/cudf/cudf/_lib/copying.pyx

@@ -384,7 +384,7 @@ cdef class _CPackedColumns:
 
         p.column_names = input_table._column_names
         p.column_dtypes = {}
-        for name, col in input_table._data.items():
+        for name, col in input_table._column_labels_and_values:
             if isinstance(col.dtype, cudf.core.dtypes._BaseDtype):
                 p.column_dtypes[name] = col.dtype
 

python/cudf/cudf/_lib/csv.pyx

@@ -273,7 +273,7 @@ def read_csv(
         elif isinstance(dtype, abc.Collection):
             for index, col_dtype in enumerate(dtype):
                 if isinstance(cudf.dtype(col_dtype), cudf.CategoricalDtype):
-                    col_name = df._data.names[index]
+                    col_name = df._column_names[index]
                     df._data[col_name] = df._data[col_name].astype(col_dtype)
 
     if names is not None and len(names) and isinstance(names[0], int):

python/cudf/cudf/_lib/io/utils.pyx

@@ -179,7 +179,7 @@ cdef update_struct_field_names(
 ):
     # Deprecated, remove in favor of add_col_struct_names
     # when a reader is ported to pylibcudf
-    for i, (name, col) in enumerate(table._data.items()):
+    for i, (name, col) in enumerate(table._column_labels_and_values):
         table._data[name] = update_column_struct_field_names(
             col, schema_info[i]
         )

python/cudf/cudf/_lib/parquet.pyx

@@ -235,24 +235,24 @@ cdef object _process_metadata(object df,
             df._index = idx
         elif set(index_col).issubset(names):
             index_data = df[index_col]
-            actual_index_names = list(index_col_names.values())
-            if len(index_data._data) == 1:
+            actual_index_names = iter(index_col_names.values())
+            if index_data._num_columns == 1:
                 idx = cudf.Index._from_column(
-                    index_data._data.columns[0],
-                    name=actual_index_names[0]
+                    index_data._columns[0],
+                    name=next(actual_index_names)
                 )
             else:
                 idx = cudf.MultiIndex.from_frame(
                     index_data,
-                    names=actual_index_names
+                    names=list(actual_index_names)
                 )
             df.drop(columns=index_col, inplace=True)
             df._index = idx
         else:
             if use_pandas_metadata:
                 df.index.names = index_col
 
-    if len(df._data.names) == 0 and column_index_type is not None:
+    if df._num_columns == 0 and column_index_type is not None:
         df._data.label_dtype = cudf.dtype(column_index_type)
 
     return df

python/cudf/cudf/_lib/utils.pyx

@@ -49,9 +49,9 @@ cdef table_view table_view_from_table(tbl, ignore_index=False) except*:
         If True, don't include the index in the columns.
     """
     return table_view_from_columns(
-        tbl._index._data.columns + tbl._data.columns
+        tbl._index._columns + tbl._columns
         if not ignore_index and tbl._index is not None
-        else tbl._data.columns
+        else tbl._columns
     )
 
 
@@ -62,7 +62,7 @@ cpdef generate_pandas_metadata(table, index):
     index_descriptors = []
     columns_to_convert = list(table._columns)
     # Columns
-    for name, col in table._data.items():
+    for name, col in table._column_labels_and_values:
         if cudf.get_option("mode.pandas_compatible"):
             # in pandas-compat mode, non-string column names are stringified.
             col_names.append(str(name))

python/cudf/cudf/core/_base_index.py

@@ -1951,7 +1951,7 @@ def drop_duplicates(
         return self._from_columns_like_self(
             drop_duplicates(
                 list(self._columns),
-                keys=range(len(self._data)),
+                keys=range(len(self._columns)),
                 keep=keep,
                 nulls_are_equal=nulls_are_equal,
             ),

python/cudf/cudf/core/column_accessor.py

@@ -151,9 +151,9 @@ def __setitem__(self, key: abc.Hashable, value: ColumnBase) -> None:
         self.set_by_label(key, value)
 
     def __delitem__(self, key: abc.Hashable) -> None:
-        old_ncols = len(self._data)
+        old_ncols = len(self)
         del self._data[key]
-        new_ncols = len(self._data)
+        new_ncols = len(self)
         self._clear_cache(old_ncols, new_ncols)
 
     def __len__(self) -> int:
@@ -213,7 +213,7 @@ def level_names(self) -> tuple[abc.Hashable, ...]:
 
     @property
     def nlevels(self) -> int:
-        if len(self._data) == 0:
+        if len(self) == 0:
             return 0
         if not self.multiindex:
             return 1
@@ -226,7 +226,7 @@ def name(self) -> abc.Hashable:
 
     @cached_property
     def nrows(self) -> int:
-        if len(self._data) == 0:
+        if len(self) == 0:
             return 0
         else:
             return len(next(iter(self.values())))
@@ -257,9 +257,9 @@ def _clear_cache(self, old_ncols: int, new_ncols: int) -> None:
         Parameters
         ----------
         old_ncols: int
-            len(self._data) before self._data was modified
+            len(self) before self._data was modified
         new_ncols: int
-            len(self._data) after self._data was modified
+            len(self) after self._data was modified
         """
         cached_properties = ("columns", "names", "_grouped_data")
         for attr in cached_properties:
@@ -335,7 +335,7 @@ def insert(
         if name in self._data:
             raise ValueError(f"Cannot insert '{name}', already exists")
 
-        old_ncols = len(self._data)
+        old_ncols = len(self)
         if loc == -1:
             loc = old_ncols
         elif not (0 <= loc <= old_ncols):
@@ -414,7 +414,7 @@ def get_labels_by_index(self, index: Any) -> tuple:
         tuple
         """
         if isinstance(index, slice):
-            start, stop, step = index.indices(len(self._data))
+            start, stop, step = index.indices(len(self))
             return self.names[start:stop:step]
         elif pd.api.types.is_integer(index):
             return (self.names[index],)
@@ -526,9 +526,9 @@ def set_by_label(self, key: abc.Hashable, value: ColumnBase) -> None:
         if len(self) > 0 and len(value) != self.nrows:
             raise ValueError("All columns must be of equal length")
 
-        old_ncols = len(self._data)
+        old_ncols = len(self)
         self._data[key] = value
-        new_ncols = len(self._data)
+        new_ncols = len(self)
         self._clear_cache(old_ncols, new_ncols)
 
     def _select_by_label_list_like(self, key: tuple) -> Self:
@@ -718,12 +718,12 @@ def droplevel(self, level: int) -> None:
         if level < 0:
             level += self.nlevels
 
-        old_ncols = len(self._data)
+        old_ncols = len(self)
         self._data = {
             _remove_key_level(key, level): value  # type: ignore[arg-type]
             for key, value in self._data.items()
         }
-        new_ncols = len(self._data)
+        new_ncols = len(self)
         self._level_names = (
             self._level_names[:level] + self._level_names[level + 1 :]
         )