Merge pull request #624 from laughingman7743/feature/remove-fastparquet-support

Remove fastparquet support

Author: laughingman7743

README.rst

@@ -62,8 +62,6 @@ Extra packages:
 +---------------+---------------------------------------+------------------+
 | Arrow         | ``pip install PyAthena[Arrow]``       | >=10.0.0         |
 +---------------+---------------------------------------+------------------+
-| fastparquet   | ``pip install PyAthena[fastparquet]`` | >=0.4.0          |
-+---------------+---------------------------------------+------------------+
 
 .. _usage:
 

docs/introduction.rst

@@ -32,8 +32,6 @@ Extra packages:
 +---------------+---------------------------------------+------------------+
 | Arrow         | ``pip install PyAthena[Arrow]``       | >=7.0.0          |
 +---------------+---------------------------------------+------------------+
-| fastparquet   | ``pip install PyAthena[fastparquet]`` | >=0.4.0          |
-+---------------+---------------------------------------+------------------+
 
 .. _features:
 

pyathena/fastparquet/__init__.py

@@ -222,13 +222,13 @@ def _get_parquet_engine(self) -> str:
         """Get the parquet engine to use, handling auto-detection.
 
         Returns:
-            Name of the parquet engine to use ('pyarrow' or 'fastparquet').
+            Name of the parquet engine to use ('pyarrow').
 
         Raises:
-            ImportError: If no suitable parquet engine is available.
+            ImportError: If pyarrow is not available.
         """
         if self._engine == "auto":
-            return self._get_available_engine(["pyarrow", "fastparquet"])
+            return self._get_available_engine(["pyarrow"])
         return self._engine
 
     def _get_csv_engine(
@@ -554,11 +554,8 @@ def _read_parquet(self, engine) -> "DataFrame":
             kwargs = {
                 "use_threads": True,
             }
-        elif engine == "fastparquet":
-            unload_location = f"{self._unload_location}*"
-            kwargs = {}
         else:
-            raise ProgrammingError("Engine must be one of `pyarrow`, `fastparquet`.")
+            raise ProgrammingError("Engine must be `pyarrow`.")
         kwargs.update(self._kwargs)
 
         try:
@@ -592,23 +589,8 @@ def _read_parquet_schema(self, engine) -> Tuple[Dict[str, Any], ...]:
             except Exception as e:
                 _logger.exception(f"Failed to read schema {bucket}/{key}.")
                 raise OperationalError(*e.args) from e
-        elif engine == "fastparquet":
-            from fastparquet import ParquetFile
-
-            # TODO: https://github.com/python/mypy/issues/1153
-            from pyathena.fastparquet.util import to_column_info  # type: ignore
-
-            if not self._data_manifest:
-                self._data_manifest = self._read_data_manifest()
-            bucket, key = parse_output_location(self._data_manifest[0])
-            try:
-                file = ParquetFile(f"{bucket}/{key}", open_with=self._fs.open)
-                return to_column_info(file.schema)
-            except Exception as e:
-                _logger.exception(f"Failed to read schema {bucket}/{key}.")
-                raise OperationalError(*e.args) from e
         else:
-            raise ProgrammingError("Engine must be one of `pyarrow`, `fastparquet`.")
+            raise ProgrammingError("Engine must be `pyarrow`.")
 
     def _as_pandas(self) -> Union["TextFileReader", "DataFrame"]:
         if self.is_unload:

pyathena/fastparquet/util.py

@@ -42,15 +42,13 @@ awsathena = "pyathena.sqlalchemy.base:AthenaDialect"
 sqlalchemy = ["sqlalchemy>=1.0.0"]
 pandas = ["pandas>=1.3.0"]
 arrow = ["pyarrow>=10.0.0"]
-fastparquet = ["fastparquet>=0.4.0"]
 
 [dependency-groups]
 dev = [
     "sqlalchemy>=1.0.0",
     "pandas>=1.3.0",
     "numpy>=1.26.0",
     "pyarrow>=10.0.0",
-    "fastparquet>=0.4.0",
     "Jinja2>=3.1.0",
     "mypy>=0.900",
     "pytest>=3.5",