REF: names no longer needed in _form_blocks (pandas-dev#43114)

Author: jbrockmendel

pandas/core/internals/__init__.py

@@ -1,4 +1,7 @@
-from pandas.core.internals.api import make_block  # pseudo-public version
+from pandas.core.internals.api import (
+    create_block_manager_from_arrays,
+    make_block,
+)
 from pandas.core.internals.array_manager import (
     ArrayManager,
     SingleArrayManager,
@@ -18,7 +21,6 @@
 from pandas.core.internals.managers import (
     BlockManager,
     SingleBlockManager,
-    create_block_manager_from_arrays,
     create_block_manager_from_blocks,
 )
 

pandas/core/internals/api.py

@@ -8,10 +8,16 @@
 """
 from __future__ import annotations
 
+from collections import defaultdict
+from typing import DefaultDict
+
 import numpy as np
 
 from pandas._libs.internals import BlockPlacement
-from pandas._typing import Dtype
+from pandas._typing import (
+    ArrayLike,
+    Dtype,
+)
 
 from pandas.core.dtypes.common import (
     is_datetime64tz_dtype,
@@ -20,14 +26,24 @@
 
 from pandas.core.arrays import DatetimeArray
 from pandas.core.construction import extract_array
+from pandas.core.indexes.api import Index
 from pandas.core.internals.blocks import (
     Block,
+    CategoricalBlock,
     DatetimeTZBlock,
+    ExtensionBlock,
     check_ndim,
     ensure_block_shape,
     extract_pandas_array,
     get_block_type,
     maybe_coerce_values,
+    new_block,
+)
+from pandas.core.internals.managers import (
+    BlockManager,
+    construction_error,
+    multi_blockify,
+    simple_blockify,
 )
 
 
@@ -86,3 +102,110 @@ def maybe_infer_ndim(values, placement: BlockPlacement, ndim: int | None) -> int
         else:
             ndim = values.ndim
     return ndim
+
+
+def create_block_manager_from_arrays(
+    arrays,
+    names: Index,
+    axes: list[Index],
+    consolidate: bool = True,
+) -> BlockManager:
+    # Assertions disabled for performance
+    # assert isinstance(names, Index)
+    # assert isinstance(axes, list)
+    # assert all(isinstance(x, Index) for x in axes)
+
+    arrays = [extract_array(x, extract_numpy=True) for x in arrays]
+
+    try:
+        blocks = _form_blocks(arrays, names, axes, consolidate)
+        mgr = BlockManager(blocks, axes)
+    except ValueError as e:
+        raise construction_error(len(arrays), arrays[0].shape, axes, e)
+    if consolidate:
+        mgr._consolidate_inplace()
+    return mgr
+
+
+def _form_blocks(
+    arrays: list[ArrayLike], names: Index, axes: list[Index], consolidate: bool
+) -> list[Block]:
+    # put "leftover" items in float bucket, where else?
+    # generalize?
+    items_dict: DefaultDict[str, list] = defaultdict(list)
+    extra_locs = []
+
+    names_idx = names
+    if names_idx.equals(axes[0]):
+        names_indexer = np.arange(len(names_idx))
+    else:
+        # Assertion disabled for performance
+        # assert names_idx.intersection(axes[0]).is_unique
+        names_indexer = names_idx.get_indexer_for(axes[0])
+
+    for i, name_idx in enumerate(names_indexer):
+        if name_idx == -1:
+            extra_locs.append(i)
+            continue
+
+        v = arrays[name_idx]
+
+        block_type = get_block_type(v)
+        items_dict[block_type.__name__].append((i, v))
+
+    blocks: list[Block] = []
+    if len(items_dict["NumericBlock"]):
+        numeric_blocks = multi_blockify(
+            items_dict["NumericBlock"], consolidate=consolidate
+        )
+        blocks.extend(numeric_blocks)
+
+    if len(items_dict["DatetimeLikeBlock"]):
+        dtlike_blocks = multi_blockify(
+            items_dict["DatetimeLikeBlock"], consolidate=consolidate
+        )
+        blocks.extend(dtlike_blocks)
+
+    if len(items_dict["DatetimeTZBlock"]):
+        dttz_blocks = [
+            DatetimeTZBlock(
+                ensure_block_shape(extract_array(array), 2),
+                placement=BlockPlacement(i),
+                ndim=2,
+            )
+            for i, array in items_dict["DatetimeTZBlock"]
+        ]
+        blocks.extend(dttz_blocks)
+
+    if len(items_dict["ObjectBlock"]) > 0:
+        object_blocks = simple_blockify(
+            items_dict["ObjectBlock"], np.object_, consolidate=consolidate
+        )
+        blocks.extend(object_blocks)
+
+    if len(items_dict["CategoricalBlock"]) > 0:
+        cat_blocks = [
+            CategoricalBlock(array, placement=BlockPlacement(i), ndim=2)
+            for i, array in items_dict["CategoricalBlock"]
+        ]
+        blocks.extend(cat_blocks)
+
+    if len(items_dict["ExtensionBlock"]):
+        external_blocks = [
+            ExtensionBlock(array, placement=BlockPlacement(i), ndim=2)
+            for i, array in items_dict["ExtensionBlock"]
+        ]
+
+        blocks.extend(external_blocks)
+
+    if len(extra_locs):
+        shape = (len(extra_locs),) + tuple(len(x) for x in axes[1:])
+
+        # empty items -> dtype object
+        block_values = np.empty(shape, dtype=object)
+        block_values.fill(np.nan)
+
+        na_block = new_block(block_values, placement=extra_locs, ndim=2)
+        blocks.append(na_block)
+
+    return blocks

pandas/core/internals/construction.py

@@ -85,8 +85,8 @@
 from pandas.core.internals.managers import (
     BlockManager,
     SingleBlockManager,
-    create_block_manager_from_arrays,
     create_block_manager_from_blocks,
+    create_block_manager_from_column_arrays,
 )
 
 if TYPE_CHECKING:
@@ -131,8 +131,8 @@ def arrays_to_mgr(
     axes = [columns, index]
 
     if typ == "block":
-        return create_block_manager_from_arrays(
-            arrays, columns, axes, consolidate=consolidate
+        return create_block_manager_from_column_arrays(
+            arrays, axes, consolidate=consolidate
         )
     elif typ == "array":
         if len(columns) != len(arrays):

pandas/core/internals/managers.py

@@ -1808,21 +1808,19 @@ def create_block_manager_from_blocks(
     return mgr
 
 
-def create_block_manager_from_arrays(
+def create_block_manager_from_column_arrays(
     arrays,
-    names: Index,
     axes: list[Index],
     consolidate: bool = True,
 ) -> BlockManager:
     # Assertions disabled for performance
-    # assert isinstance(names, Index)
     # assert isinstance(axes, list)
     # assert all(isinstance(x, Index) for x in axes)
 
     arrays = [extract_array(x, extract_numpy=True) for x in arrays]
 
     try:
-        blocks = _form_blocks(arrays, names, axes, consolidate)
+        blocks = _form_blocks(arrays, consolidate)
         mgr = BlockManager(blocks, axes)
     except ValueError as e:
         raise construction_error(len(arrays), arrays[0].shape, axes, e)
@@ -1860,26 +1858,11 @@ def construction_error(
 # -----------------------------------------------------------------------
 
 
-def _form_blocks(
-    arrays: list[ArrayLike], names: Index, axes: list[Index], consolidate: bool
-) -> list[Block]:
-    # put "leftover" items in float bucket, where else?
-    # generalize?
-    items_dict: DefaultDict[str, list] = defaultdict(list)
-    extra_locs = []
+def _form_blocks(arrays: list[ArrayLike], consolidate: bool) -> list[Block]:
 
-    names_idx = names
-    if names_idx.equals(axes[0]):
-        names_indexer = np.arange(len(names_idx))
-    else:
-        # Assertion disabled for performance
-        # assert names_idx.intersection(axes[0]).is_unique
-        names_indexer = names_idx.get_indexer_for(axes[0])
+    items_dict: DefaultDict[str, list] = defaultdict(list)
 
-    for i, name_idx in enumerate(names_indexer):
-        if name_idx == -1:
-            extra_locs.append(i)
-            continue
+    for i, name_idx in enumerate(range(len(arrays))):
 
         v = arrays[name_idx]
 
@@ -1888,13 +1871,13 @@ def _form_blocks(
 
     blocks: list[Block] = []
     if len(items_dict["NumericBlock"]):
-        numeric_blocks = _multi_blockify(
+        numeric_blocks = multi_blockify(
             items_dict["NumericBlock"], consolidate=consolidate
         )
         blocks.extend(numeric_blocks)
 
     if len(items_dict["DatetimeLikeBlock"]):
-        dtlike_blocks = _multi_blockify(
+        dtlike_blocks = multi_blockify(
             items_dict["DatetimeLikeBlock"], consolidate=consolidate
         )
         blocks.extend(dtlike_blocks)
@@ -1911,7 +1894,7 @@ def _form_blocks(
         blocks.extend(dttz_blocks)
 
     if len(items_dict["ObjectBlock"]) > 0:
-        object_blocks = _simple_blockify(
+        object_blocks = simple_blockify(
             items_dict["ObjectBlock"], np.object_, consolidate=consolidate
         )
         blocks.extend(object_blocks)
@@ -1931,20 +1914,10 @@ def _form_blocks(
 
         blocks.extend(external_blocks)
 
-    if len(extra_locs):
-        shape = (len(extra_locs),) + tuple(len(x) for x in axes[1:])
-
-        # empty items -> dtype object
-        block_values = np.empty(shape, dtype=object)
-        block_values.fill(np.nan)
-
-        na_block = new_block(block_values, placement=extra_locs, ndim=2)
-        blocks.append(na_block)
-
     return blocks
 
 
-def _simple_blockify(tuples, dtype, consolidate: bool) -> list[Block]:
+def simple_blockify(tuples, dtype, consolidate: bool) -> list[Block]:
     """
     return a single array of a block that has a single dtype; if dtype is
     not None, coerce to this dtype
@@ -1962,7 +1935,7 @@ def _simple_blockify(tuples, dtype, consolidate: bool) -> list[Block]:
     return [block]
 
 
-def _multi_blockify(tuples, dtype: DtypeObj | None = None, consolidate: bool = True):
+def multi_blockify(tuples, dtype: DtypeObj | None = None, consolidate: bool = True):
     """return an array of blocks that potentially have different dtypes"""
 
     if not consolidate: