maxframe 0.1.0b2__cp310-cp310-win_amd64.whl → 0.1.0b4__cp310-cp310-win_amd64.whl

maxframe/codegen.py

@@ -17,7 +17,7 @@ import base64
 import dataclasses
 import logging
 from enum import Enum
-from typing import TYPE_CHECKING, Any, Dict, List, Optional, Tuple, Type
+from typing import TYPE_CHECKING, Any, Dict, List, Optional, Tuple, Type, Union
 
 from odps.types import OdpsSchema
 from odps.utils import camel_to_underline
@@ -30,6 +30,7 @@ from .odpsio import build_dataframe_table_meta
 from .odpsio.schema import pandas_to_odps_schema
 from .protocol import DataFrameTableMeta, ResultInfo
 from .serialization import PickleContainer
+from .serialization.serializables import Serializable, StringField
 from .typing_ import PandasObjectTypes
 from .udf import MarkedFunction
 
@@ -48,8 +49,11 @@ class CodeGenResult:
     constants: Dict[str, Any]
 
 
-class AbstractUDF(abc.ABC):
-    _session_id: str
+class AbstractUDF(Serializable):
+    _session_id: str = StringField("session_id")
+
+    def __init__(self, session_id: Optional[str] = None, **kw):
+        super().__init__(_session_id=session_id, **kw)
 
     @property
     def name(self) -> str:
@@ -74,7 +78,66 @@ class AbstractUDF(abc.ABC):
 
 class UserCodeMixin:
     @classmethod
-    def generate_pickled_codes(cls, code_to_pickle: Any) -> List[str]:
+    def obj_to_python_expr(cls, obj: Any = None) -> str:
+        """
+        Parameters
+        ----------
+        obj
+            The object to convert to python expr.
+        Returns
+        -------
+        str :
+            The str type content equals to the object when use in the python code directly.
+        """
+        if obj is None:
+            return "None"
+
+        if isinstance(obj, (int, float)):
+            return repr(obj)
+
+        if isinstance(obj, bool):
+            return "True" if obj else "False"
+
+        if isinstance(obj, bytes):
+            base64_bytes = base64.b64encode(obj)
+            return f"base64.b64decode({base64_bytes})"
+
+        if isinstance(obj, str):
+            return repr(obj)
+
+        if isinstance(obj, list):
+            return (
+                f"[{', '.join([cls.obj_to_python_expr(element) for element in obj])}]"
+            )
+
+        if isinstance(obj, dict):
+            items = (
+                f"{repr(key)}: {cls.obj_to_python_expr(value)}"
+                for key, value in obj.items()
+            )
+            return f"{{{', '.join(items)}}}"
+
+        if isinstance(obj, tuple):
+            return f"({', '.join([cls.obj_to_python_expr(sub_obj) for sub_obj in obj])}{',' if len(obj) == 1 else ''})"
+
+        if isinstance(obj, set):
+            return (
+                f"{{{', '.join([cls.obj_to_python_expr(sub_obj) for sub_obj in obj])}}}"
+                if obj
+                else "set()"
+            )
+
+        if isinstance(obj, PickleContainer):
+            return UserCodeMixin.generate_pickled_codes(obj, None)
+
+        raise ValueError(f"not support arg type {type(obj)}")
+
+    @classmethod
+    def generate_pickled_codes(
+        cls,
+        code_to_pickle: Any,
+        unpicked_data_var_name: Union[str, None] = "pickled_data",
+    ) -> str:
         """
         Generate pickled codes. The final pickled variable is called 'pickled_data'.
 
@@ -82,20 +145,20 @@ class UserCodeMixin:
         ----------
         code_to_pickle: Any
             The code to be pickled.
+        unpicked_data_var_name: str
+            The variables in code used to hold the loads object from the cloudpickle
 
         Returns
         -------
-        List[str] :
-            The code snippets of pickling, the final variable is called 'pickled_data'.
+        str :
+            The code snippets of pickling, the final variable is called 'pickled_data' by default.
         """
         pickled, buffers = cls.dump_pickled_data(code_to_pickle)
-        pickled = base64.b64encode(pickled)
-        buffers = [base64.b64encode(b) for b in buffers]
-        buffers_str = ", ".join(f"base64.b64decode(b'{b.decode()}')" for b in buffers)
-        return [
-            f"base64_data = base64.b64decode(b'{pickled.decode()}')",
-            f"pickled_data = cloudpickle.loads(base64_data, buffers=[{buffers_str}])",
-        ]
+        pickle_loads_expr = f"cloudpickle.loads({cls.obj_to_python_expr(pickled)}, buffers={cls.obj_to_python_expr(buffers)})"
+        if unpicked_data_var_name:
+            return f"{unpicked_data_var_name} = {pickle_loads_expr}"
+
+        return pickle_loads_expr
 
     @staticmethod
     def dump_pickled_data(
@@ -114,8 +177,9 @@ class UserCodeMixin:
 
 
 class BigDagCodeContext(metaclass=abc.ABCMeta):
-    def __init__(self, session_id: str = None):
+    def __init__(self, session_id: str = None, subdag_id: str = None):
         self._session_id = session_id
+        self._subdag_id = subdag_id
         self._tileable_key_to_variables = dict()
         self.constants = dict()
         self._data_table_meta_cache = dict()
@@ -142,10 +206,14 @@ class BigDagCodeContext(metaclass=abc.ABCMeta):
         except KeyError:
             var_name = self._tileable_key_to_variables[
                 tileable.key
-            ] = f"var_{self._next_var_id}"
-            self._next_var_id += 1
+            ] = self.next_var_name()
             return var_name
 
+    def next_var_name(self) -> str:
+        var_name = f"var_{self._next_var_id}"
+        self._next_var_id += 1
+        return var_name
+
     def get_odps_schema(
         self, data: PandasObjectTypes, unknown_as_string: bool = False
     ) -> OdpsSchema:
@@ -275,9 +343,10 @@ class BigDagCodeGenerator(metaclass=abc.ABCMeta):
     engine_priority: int = 0
     _extension_loaded = False
 
-    def __init__(self, session_id: str):
+    def __init__(self, session_id: str, subdag_id: str = None):
         self._session_id = session_id
-        self._context = self._init_context(session_id)
+        self._subdag_id = subdag_id
+        self._context = self._init_context(session_id, subdag_id)
 
     @classmethod
     def _load_engine_extensions(cls):
@@ -307,7 +376,7 @@ class BigDagCodeGenerator(metaclass=abc.ABCMeta):
         raise NotImplementedError
 
     @abc.abstractmethod
-    def _init_context(self, session_id: str) -> BigDagCodeContext:
+    def _init_context(self, session_id: str, subdag_id: str) -> BigDagCodeContext:
         raise NotImplementedError
 
     def _generate_comments(

maxframe/config/config.py

@@ -340,6 +340,12 @@ default_options.register_option(
     validator=is_integer,
     remote=True,
 )
+default_options.register_option(
+    "session.subinstance_priority",
+    None,
+    validator=any_validator(is_null, is_integer),
+    remote=True,
+)
 
 default_options.register_option("warn_duplicated_execution", False, validator=is_bool)
 default_options.register_option("dataframe.use_arrow_dtype", True, validator=is_bool)
@@ -352,6 +358,9 @@ default_options.register_option(
 default_options.register_option(
     "show_progress", "auto", validator=any_validator(is_bool, is_string)
 )
+default_options.register_option(
+    "dag.settings", value=dict(), validator=is_dict, remote=True
+)
 
 ################
 # SPE Settings #

maxframe/core/entity/executable.py

@@ -66,6 +66,7 @@ class DecrefRunner:
         if self._decref_thread:  # pragma: no branch
             self._queue.put_nowait((None, None, None))
             self._decref_thread.join(1)
+        self._decref_thread = None
 
     def put(self, key: str, session_ref: ref):
         if self._decref_thread is None:

maxframe/core/entity/objects.py

@@ -15,6 +15,7 @@
 from typing import Any, Dict
 
 from ...serialization.serializables import FieldTypes, ListField
+from ...utils import skip_na_call
 from .chunks import Chunk, ChunkData
 from .core import Entity
 from .executable import _ToObjectMixin
@@ -62,8 +63,8 @@ class ObjectData(TileableData, _ToObjectMixin):
     _chunks = ListField(
         "chunks",
         FieldTypes.reference(ObjectChunkData),
-        on_serialize=lambda x: [it.data for it in x] if x is not None else x,
-        on_deserialize=lambda x: [ObjectChunk(it) for it in x] if x is not None else x,
+        on_serialize=skip_na_call(lambda x: [it.data for it in x]),
+        on_deserialize=skip_na_call(lambda x: [ObjectChunk(it) for it in x]),
     )
 
     def __init__(self, op=None, nsplits=None, **kw):

maxframe/dataframe/__init__.py

@@ -39,6 +39,7 @@ from .datasource.read_odps_query import read_odps_query
 from .datasource.read_odps_table import read_odps_table
 from .datasource.read_parquet import read_parquet
 from .datastore.to_odps import to_odps_table
+from .groupby import NamedAgg
 from .initializer import DataFrame, Index, Series, read_pandas
 from .merge import concat, merge
 from .misc.cut import cut
@@ -52,7 +53,12 @@ from .reduction import CustomReduction, unique
 from .tseries.to_datetime import to_datetime
 
 try:
-    from pandas import NA, NamedAgg, Timestamp
+    from pandas import NA, Timestamp
+except ImportError:  # pragma: no cover
+    pass
+
+try:
+    from . import _internal
 except ImportError:  # pragma: no cover
     pass
 

maxframe/dataframe/core.py

@@ -960,7 +960,9 @@ class BaseSeriesData(HasShapeTileableData, _ToPandasMixin):
             buf = StringIO()
             max_rows = pd.get_option("display.max_rows")
             corner_max_rows = (
-                max_rows if self.shape[0] <= max_rows else corner_data.shape[0] - 1
+                max_rows
+                if self.shape[0] <= max_rows or corner_data.shape[0] == 0
+                else corner_data.shape[0] - 1
             )  # make sure max_rows < corner_data
 
             with pd.option_context("display.max_rows", corner_max_rows):
@@ -1605,7 +1607,7 @@ class DataFrameData(_BatchedFetcher, BaseDataFrameData):
             buf = StringIO()
             max_rows = pd.get_option("display.max_rows")
 
-            if self.shape[0] <= max_rows:
+            if self.shape[0] <= max_rows or corner_data.shape[0] == 0:
                 buf.write(repr(corner_data) if representation else str(corner_data))
             else:
                 # remember we cannot directly call repr(df),

maxframe/dataframe/datasource/read_odps_query.py

@@ -46,7 +46,7 @@ _EXPLAIN_TASK_SCHEMA_REGEX = re.compile(
     r"In Task ([^:]+)[\S\s]+FS: output: ([^\n #]+)[\s\S]+schema:\s+([\S\s]+)$",
     re.MULTILINE,
 )
-_EXPLAIN_COLUMN_REGEX = re.compile(r"([^ ]+) \(([^)]+)\)(?:| AS ([^ ]+))(?:\n|$)")
+_EXPLAIN_COLUMN_REGEX = re.compile(r"([^\(]+) \(([^)]+)\)(?:| AS ([^ ]+))(?:\n|$)")
 
 
 @dataclasses.dataclass
@@ -263,7 +263,9 @@ def read_odps_query(
     result: DataFrame
         DataFrame read from MaxCompute (ODPS) table
     """
-    odps_entry = odps_entry or ODPS.from_environments()
+    odps_entry = odps_entry or ODPS.from_global() or ODPS.from_environments()
+    if odps_entry is None:
+        raise ValueError("Missing odps_entry parameter")
     inst = odps_entry.execute_sql(f"EXPLAIN {query}")
     explain_str = list(inst.get_task_results().values())[0]
 

maxframe/dataframe/datasource/read_odps_table.py

@@ -69,7 +69,7 @@ class DataFrameReadODPSTable(
         return getattr(self, "partition_spec", None)
 
     def get_columns(self):
-        return self.columns
+        return self.columns or list(self.dtypes.index)
 
     def set_pruned_columns(self, columns, *, keep_order=None):  # pragma: no cover
         self.columns = columns
@@ -164,6 +164,8 @@ def read_odps_table(
         DataFrame read from MaxCompute (ODPS) table
     """
     odps_entry = odps_entry or ODPS.from_global() or ODPS.from_environments()
+    if odps_entry is None:
+        raise ValueError("Missing odps_entry parameter")
     if isinstance(table_name, Table):
         table = table_name
     else:

maxframe/dataframe/datasource/tests/test_datasource.py

@@ -12,6 +12,7 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
+import os
 from collections import OrderedDict
 
 import numpy as np
@@ -33,6 +34,7 @@ from ..from_tensor import (
 )
 from ..index import from_pandas as from_pandas_index
 from ..index import from_tileable
+from ..read_odps_query import ColumnSchema, _resolve_task_sector
 from ..series import from_pandas as from_pandas_series
 
 ray = lazy_import("ray")
@@ -228,6 +230,7 @@ def test_from_odps_table():
     assert df.op.table_name == test_table.full_table_name
     assert df.index_value.name is None
     assert isinstance(df.index_value.value, IndexValue.RangeIndex)
+    assert df.op.get_columns() == ["col1", "col2", "col3"]
     pd.testing.assert_series_equal(
         df.dtypes,
         pd.Series(
@@ -247,6 +250,7 @@ def test_from_odps_table():
     assert df.op.table_name == test_table.full_table_name
     assert df.index_value.name is None
     assert isinstance(df.index_value.value, IndexValue.RangeIndex)
+    assert df.op.get_columns() == ["col1", "col2"]
     pd.testing.assert_series_equal(
         df.dtypes,
         pd.Series([np.dtype("O"), np.dtype("int64")], index=["col1", "col2"]),
@@ -257,6 +261,7 @@ def test_from_odps_table():
     assert df.index_value.name == "col1"
     assert isinstance(df.index_value.value, IndexValue.Index)
     assert df.index.dtype == np.dtype("O")
+    assert df.op.get_columns() == ["col2", "col3"]
     pd.testing.assert_series_equal(
         df.dtypes,
         pd.Series([np.dtype("int64"), np.dtype("float64")], index=["col2", "col3"]),
@@ -267,6 +272,7 @@ def test_from_odps_table():
 
     df = read_odps_table(test_parted_table, append_partitions=True)
     assert df.op.append_partitions is True
+    assert df.op.get_columns() == ["col1", "col2", "col3", "pt"]
     pd.testing.assert_series_equal(
         df.dtypes,
         pd.Series(
@@ -280,6 +286,7 @@ def test_from_odps_table():
     )
     assert df.op.append_partitions is True
     assert df.op.partitions == ["pt=20240103"]
+    assert df.op.get_columns() == ["col1", "col2", "pt"]
     pd.testing.assert_series_equal(
         df.dtypes,
         pd.Series(
@@ -377,3 +384,18 @@ def test_date_range():
     assert dr.index_value.is_unique == expected.is_unique
     assert dr.index_value.is_monotonic_increasing == expected.is_monotonic_increasing
     assert dr.name == expected.name
+
+
+def test_resolve_task_sector():
+    input_path = os.path.join(os.path.dirname(__file__), "test-data", "task-input.txt")
+    with open(input_path, "r") as f:
+        sector = f.read()
+    actual_sector = _resolve_task_sector("job0", sector)
+
+    assert actual_sector.job_name == "job0"
+    assert actual_sector.task_name == "M1"
+    assert actual_sector.output_target == "Screen"
+    assert len(actual_sector.schema) == 78
+    assert actual_sector.schema[0] == ColumnSchema("unnamed: 0", "bigint", "")
+    assert actual_sector.schema[1] == ColumnSchema("id", "bigint", "id_alias")
+    assert actual_sector.schema[2] == ColumnSchema("listing_url", "string", "")

maxframe/dataframe/datastore/core.py

@@ -0,0 +1,19 @@
+# Copyright 1999-2024 Alibaba Group Holding Ltd.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#      http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+from ..operators import DataFrameOperator, DataFrameOperatorMixin
+
+
+class DataFrameDataStore(DataFrameOperator, DataFrameOperatorMixin):
+    pass

maxframe/dataframe/datastore/to_csv.py

@@ -23,11 +23,11 @@ from ...serialization.serializables import (
     ListField,
     StringField,
 )
-from ..operators import DataFrameOperator, DataFrameOperatorMixin
 from ..utils import parse_index
+from .core import DataFrameDataStore
 
 
-class DataFrameToCSV(DataFrameOperator, DataFrameOperatorMixin):
+class DataFrameToCSV(DataFrameDataStore):
     _op_type_ = opcodes.TO_CSV
 
     input = KeyField("input")

maxframe/dataframe/datastore/to_odps.py

@@ -32,13 +32,13 @@ from ...serialization.serializables import (
 )
 from ...typing_ import TileableType
 from ..core import DataFrame  # noqa: F401
-from ..operators import DataFrameOperator, DataFrameOperatorMixin
 from ..utils import parse_index
+from .core import DataFrameDataStore
 
 logger = logging.getLogger(__name__)
 
 
-class DataFrameToODPSTable(DataFrameOperator, DataFrameOperatorMixin):
+class DataFrameToODPSTable(DataFrameDataStore):
     _op_type_ = opcodes.TO_ODPS_TABLE
 
     dtypes = SeriesField("dtypes")

maxframe/dataframe/groupby/__init__.py

@@ -14,6 +14,7 @@
 
 # noinspection PyUnresolvedReferences
 from ..core import DataFrameGroupBy, GroupBy, SeriesGroupBy
+from .core import NamedAgg
 
 
 def _install():

maxframe/dataframe/groupby/core.py

@@ -12,6 +12,8 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
+from collections import namedtuple
+
 import pandas as pd
 
 from ... import opcodes
@@ -30,6 +32,9 @@ _GROUP_KEYS_NO_DEFAULT = pd_release_version >= (1, 5, 0)
 _default_group_keys = no_default if _GROUP_KEYS_NO_DEFAULT else True
 
 
+NamedAgg = namedtuple("NamedAgg", ["column", "aggfunc"])
+
+
 class DataFrameGroupByOperator(MapReduceOperator, DataFrameOperatorMixin):
     _op_type_ = opcodes.GROUPBY
 

maxframe/dataframe/indexing/reset_index.py

@@ -107,7 +107,6 @@ def df_reset_index(
     inplace=False,
     col_level=0,
     col_fill="",
-    incremental_index=False,
 ):
     """
     Reset the index, or a level of it.
@@ -133,12 +132,6 @@ def df_reset_index(
     col_fill : object, default ''
         If the columns have multiple levels, determines how the other
         levels are named. If None then the index name is repeated.
-    incremental_index: bool, default False
-        Ensure RangeIndex incremental, when output DataFrame has multiple chunks,
-        ensuring index incremental costs more computation,
-        so by default, each chunk will have index which starts from 0,
-        setting incremental_index=True，reset_index will guarantee that
-        output DataFrame's index is from 0 to n - 1.
 
     Returns
     -------
@@ -264,7 +257,6 @@ def df_reset_index(
         drop=drop,
         col_level=col_level,
         col_fill=col_fill,
-        incremental_index=incremental_index,
         output_types=[OutputType.dataframe],
     )
     ret = op(df)
@@ -280,7 +272,6 @@ def series_reset_index(
     drop=False,
     name=no_default,
     inplace=False,
-    incremental_index=False,
 ):
     """
     Generate a new DataFrame or Series with the index reset.
@@ -303,12 +294,6 @@ def series_reset_index(
         when `drop` is True.
     inplace : bool, default False
         Modify the Series in place (do not create a new object).
-    incremental_index: bool, default False
-        Ensure RangeIndex incremental, when output Series has multiple chunks,
-        ensuring index incremental costs more computation,
-        so by default, each chunk will have index which starts from 0,
-        setting incremental_index=True，reset_index will guarantee that
-        output Series's index is from 0 to n - 1.
 
     Returns
     -------
@@ -406,8 +391,7 @@ def series_reset_index(
         level=level,
         drop=drop,
         name=name,
-        incremental_index=incremental_index,
-        output_types=[OutputType.series],
+        output_types=[OutputType.series if drop else OutputType.dataframe],
     )
     ret = op(series)
     if not inplace:

maxframe/lib/aio/isolation.py

@@ -14,11 +14,14 @@
 
 import asyncio
 import atexit
+import itertools
 import threading
 from typing import Dict, Optional
 
 
 class Isolation:
+    _counter = itertools.count().__next__
+
     loop: asyncio.AbstractEventLoop
     _stopped: Optional[asyncio.Event]
     _thread: Optional[threading.Thread]
@@ -38,7 +41,9 @@ class Isolation:
 
     def start(self):
         if self._threaded:
-            self._thread = thread = threading.Thread(target=self._run)
+            self._thread = thread = threading.Thread(
+                name=f"IsolationThread-{self._counter()}", target=self._run
+            )
             thread.daemon = True
             thread.start()
             self._thread_ident = thread.ident

maxframe/odpsio/arrow.py

@@ -65,14 +65,19 @@ def arrow_to_pandas(
         raise ValueError(f"Does not support meta type {table_meta.type!r}")
 
 
-def pandas_to_arrow(df: Any, nthreads=1) -> Tuple[ArrowTableType, DataFrameTableMeta]:
-    table_meta = build_dataframe_table_meta(df)
+def pandas_to_arrow(
+    df: Any, nthreads=1, ignore_index=False
+) -> Tuple[ArrowTableType, DataFrameTableMeta]:
+    table_meta = build_dataframe_table_meta(df, ignore_index)
     df = df.copy() if callable(getattr(df, "copy", None)) else df
     if table_meta.type in (OutputType.dataframe, OutputType.series):
         if table_meta.type == OutputType.series:
             df = df.to_frame("_data" if df.name is None else df.name)
         df.columns = pd.Index(table_meta.table_column_names)
-        df = df.rename_axis(table_meta.table_index_column_names).reset_index()
+        if not ignore_index:
+            df = df.rename_axis(table_meta.table_index_column_names).reset_index()
+    elif ignore_index:
+        df = pd.DataFrame([], columns=[])
     elif table_meta.type == OutputType.index:
         names = [f"_idx_{idx}" for idx in range(len(df.names))]
         df = df.to_frame(name=names[0] if len(names) == 1 else names)

maxframe/odpsio/schema.py

@@ -175,7 +175,9 @@ def _scalar_as_index(df_obj: Any) -> pd.Index:
 
 
 def pandas_to_odps_schema(
-    df_obj: Any, unknown_as_string: bool = False
+    df_obj: Any,
+    unknown_as_string: bool = False,
+    ignore_index=False,
 ) -> Tuple[odps_types.OdpsSchema, DataFrameTableMeta]:
     from .. import dataframe as md
     from .arrow import pandas_to_arrow
@@ -209,7 +211,7 @@ def pandas_to_odps_schema(
     else:
         empty_df_obj = df_obj
 
-    arrow_data, table_meta = pandas_to_arrow(empty_df_obj)
+    arrow_data, table_meta = pandas_to_arrow(empty_df_obj, ignore_index=ignore_index)
     return (
         arrow_schema_to_odps_schema(
             arrow_data.schema, unknown_as_string=unknown_as_string
@@ -268,7 +270,9 @@ def build_table_column_name(
     return col_name
 
 
-def build_dataframe_table_meta(df_obj: Any) -> DataFrameTableMeta:
+def build_dataframe_table_meta(
+    df_obj: Any, ignore_index: bool = False
+) -> DataFrameTableMeta:
     from .. import dataframe as md
 
     col_to_count = defaultdict(lambda: 0)
@@ -285,6 +289,8 @@ def build_dataframe_table_meta(df_obj: Any) -> DataFrameTableMeta:
     else:  # pragma: no cover
         raise TypeError(f"Cannot accept type {type(df_obj)}")
 
+    assert not ignore_index or obj_type in (OutputType.dataframe, OutputType.series)
+
     if obj_type == OutputType.scalar:
         pd_dtypes = pd.Series([])
         column_index_names = []
@@ -340,12 +346,19 @@ def build_dataframe_table_meta(df_obj: Any) -> DataFrameTableMeta:
     else:
         index_dtypes = pd.Series([pd_index_val.dtype], index=pd_index_val.names)
 
+    if ignore_index:
+        table_index_column_names = []
+        pd_index_dtypes = pd.Series([], index=[])
+    else:
+        table_index_column_names = [f"_idx_{i}" for i in range(len(index_obj.names))]
+        pd_index_dtypes = index_dtypes
+
     return DataFrameTableMeta(
         table_name=table_name,
         type=obj_type,
         table_column_names=final_sql_columns,
-        table_index_column_names=[f"_idx_{i}" for i in range(len(index_obj.names))],
+        table_index_column_names=table_index_column_names,
         pd_column_dtypes=pd_dtypes,
         pd_column_level_names=column_index_names,
-        pd_index_dtypes=index_dtypes,
+        pd_index_dtypes=pd_index_dtypes,
     )