From 086a36da66292112ec66dad24a43ebd419cc5184 Mon Sep 17 00:00:00 2001
From: akashthemosh <107898164+akashthemosh@users.noreply.github.com>
Date: Wed, 12 Jun 2024 03:54:11 +0000
Subject: [PATCH] Fix: Add to_pandas_kwargs to read_parquet for PyArrow engine

Adds the `to_pandas_kwargs` parameter to `pd.read_parquet` to allow passing arguments to `pyarrow.Table.to_pandas`. This addresses issues that may arise during Parquet-to-DataFrame conversion, such as handling microsecond timestamps.

Fixes #49236
---
 pandas/io/parquet.py | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)

diff --git a/pandas/io/parquet.py b/pandas/io/parquet.py
index 306b144811898..0fb6f05b8b471 100644
--- a/pandas/io/parquet.py
+++ b/pandas/io/parquet.py
@@ -245,6 +245,7 @@ def read(
         dtype_backend: DtypeBackend | lib.NoDefault = lib.no_default,
         storage_options: StorageOptions | None = None,
         filesystem=None,
+        to_pandas_kwargs: dict[str, Any] | None = None,
         **kwargs,
     ) -> DataFrame:
         kwargs["use_pandas_metadata"] = True
@@ -280,7 +281,7 @@ def read(
                     "make_block is deprecated",
                     DeprecationWarning,
                 )
-                result = pa_table.to_pandas(**to_pandas_kwargs)
+                result = pa_table.to_pandas(**(to_pandas_kwargs or {}))
 
             if pa_table.schema.metadata:
                 if b"PANDAS_ATTRS" in pa_table.schema.metadata:
@@ -505,6 +506,7 @@ def read_parquet(
     dtype_backend: DtypeBackend | lib.NoDefault = lib.no_default,
     filesystem: Any = None,
     filters: list[tuple] | list[list[tuple]] | None = None,
+    to_pandas_kwargs: dict[str, Any] | None = None,
     **kwargs,
 ) -> DataFrame:
     """
@@ -540,6 +542,11 @@ def read_parquet(
         If not None, only these columns will be read from the file.
     {storage_options}
 
+     to_pandas_kwargs : dict, default None
+     Additional keyword arguments passed to :meth:`pyarrow.Table.to_pandas`
+       to control how the pyarrow Table is converted to a pandas DataFrame.
+   This is only used when `engine="pyarrow"`.
+
         .. versionadded:: 1.3.0
 
     dtype_backend : {{'numpy_nullable', 'pyarrow'}}, default 'numpy_nullable'
@@ -649,5 +656,6 @@ def read_parquet(
         storage_options=storage_options,
         dtype_backend=dtype_backend,
         filesystem=filesystem,
+        to_pandas_kwargs=to_pandas_kwargs,
         **kwargs,
     )