zhengruifeng commented on code in PR #53718: URL: https://github.com/apache/spark/pull/53718#discussion_r2670503818
########## python/pyspark/tests/upstream/pyarrow/test_pyarrow_type_inference.py: ########## @@ -0,0 +1,529 @@ +# +# Licensed to the Apache Software Foundation (ASF) under one or more +# contributor license agreements. See the NOTICE file distributed with +# this work for additional information regarding copyright ownership. +# The ASF licenses this file to You 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. +# + +import datetime +from decimal import Decimal +import unittest + +from pyspark.testing.utils import ( + have_pandas, + have_pyarrow, + pandas_requirement_message, + pyarrow_requirement_message, +) + + +# Test pa.array type inference for various input types. +# This test monitors the behavior of PyArrow's type inference to ensure +# PySpark's assumptions about PyArrow behavior remain valid across versions. +# +# Key input types tested: +# 1. nullable data (with None values) +# 2. plain Python instances (list, tuple, array) +# 3. pandas instances (pd.Series) +# 4. numpy instances (np.array) + + [email protected](not have_pyarrow, pyarrow_requirement_message) +class PyArrowTypeInferenceTests(unittest.TestCase): + """Test PyArrow's type inference behavior for pa.array.""" + + def test_nullable_data(self): + """Test type inference with nullable data (None values) for all types.""" + import pyarrow as pa + + # Single None value + a = pa.array([None]) + self.assertEqual(a.type, pa.null()) + + # Integers with None + a = pa.array([1, 2, 3]) + self.assertEqual(a.type, pa.int64()) + + a = pa.array([1, 2, 3, None]) + self.assertEqual(a.type, pa.int64()) + + # Floats with None + a = pa.array([1.0, 2.0, 3.0]) + self.assertEqual(a.type, pa.float64()) + + a = pa.array([1.0, 2.0, None]) + self.assertEqual(a.type, pa.float64()) + + # Strings with None + a = pa.array(["a", "b", "c"]) + self.assertEqual(a.type, pa.string()) + + a = pa.array(["a", "b", None]) + self.assertEqual(a.type, pa.string()) + + # Booleans with None + a = pa.array([True, False, None]) + self.assertEqual(a.type, pa.bool_()) + + # Date with None + a = pa.array([datetime.date(2024, 1, 1), None]) + self.assertEqual(a.type, pa.date32()) + + # Datetime with None + a = pa.array([datetime.datetime(2024, 1, 1, 12, 0, 0), None]) + self.assertEqual(a.type, pa.timestamp("us")) + + # Time with None + a = pa.array([datetime.time(12, 30, 0), None]) + self.assertEqual(a.type, pa.time64("us")) + + # Timedelta with None + a = pa.array([datetime.timedelta(days=1), None]) + self.assertEqual(a.type, pa.duration("us")) + + # Binary with None + a = pa.array([b"hello", None]) + self.assertEqual(a.type, pa.binary()) + + # Decimal with None + a = pa.array([Decimal("1.23"), None]) + self.assertEqual(a.type, pa.decimal128(3, 2)) + + def test_plain_python_list(self): + """Test type inference from Python lists for all types.""" + import pyarrow as pa + + # Integer list + a = pa.array([1, 2, 3]) + self.assertEqual(a.type, pa.int64()) + self.assertEqual(a.to_pylist(), [1, 2, 3]) + + # Float list + a = pa.array([1.0, 2.0, 3.0]) + self.assertEqual(a.type, pa.float64()) + + # Mixed int and float + a = pa.array([1, 2.0, 3]) + self.assertEqual(a.type, pa.float64()) + + # String list + a = pa.array(["a", "b", "c"]) + self.assertEqual(a.type, pa.string()) + + # Boolean list + a = pa.array([True, False, True]) + self.assertEqual(a.type, pa.bool_()) + + # Nested list (list of lists) + a = pa.array([[1, 2], [3, 4]]) + self.assertEqual(a.type, pa.list_(pa.int64())) + + # Date list + a = pa.array([datetime.date(2024, 1, 1), datetime.date(2024, 1, 2)]) + self.assertEqual(a.type, pa.date32()) + + # Datetime list (naive - no timezone) + a = pa.array([datetime.datetime(2024, 1, 1, 12, 0, 0), datetime.datetime(2024, 1, 2, 12, 0, 0)]) + self.assertEqual(a.type, pa.timestamp("us")) + + # Time list + a = pa.array([datetime.time(12, 30, 0), datetime.time(13, 45, 0)]) + self.assertEqual(a.type, pa.time64("us")) + + # Timedelta list (duration) + a = pa.array([datetime.timedelta(days=1), datetime.timedelta(hours=2)]) + self.assertEqual(a.type, pa.duration("us")) + + # Binary list + a = pa.array([b"hello", b"world"]) + self.assertEqual(a.type, pa.binary()) + + # Decimal list + a = pa.array([Decimal("1.23"), Decimal("4.56")]) + self.assertEqual(a.type, pa.decimal128(3, 2)) + + a = pa.array([Decimal("123456.789"), Decimal("987654.321")]) + self.assertEqual(a.type, pa.decimal128(9, 3)) + + def test_plain_python_tuple(self): + """Test type inference from Python tuples for all types.""" + import pyarrow as pa + + # Tuple of integers + a = pa.array((1, 2, 3)) + self.assertEqual(a.type, pa.int64()) + self.assertEqual(a.to_pylist(), [1, 2, 3]) + + # Tuple of floats + a = pa.array((1.0, 2.0, 3.0)) + self.assertEqual(a.type, pa.float64()) + + # Tuple of strings + a = pa.array(("a", "b", "c")) + self.assertEqual(a.type, pa.string()) + + # Tuple of booleans + a = pa.array((True, False, True)) + self.assertEqual(a.type, pa.bool_()) + + # Tuple of dates + a = pa.array((datetime.date(2024, 1, 1), datetime.date(2024, 1, 2))) + self.assertEqual(a.type, pa.date32()) + + # Tuple of datetimes + a = pa.array((datetime.datetime(2024, 1, 1, 12, 0, 0), datetime.datetime(2024, 1, 2, 12, 0, 0))) + self.assertEqual(a.type, pa.timestamp("us")) + + # Tuple of times + a = pa.array((datetime.time(12, 30, 0), datetime.time(13, 45, 0))) + self.assertEqual(a.type, pa.time64("us")) + + # Tuple of timedeltas + a = pa.array((datetime.timedelta(days=1), datetime.timedelta(hours=2))) + self.assertEqual(a.type, pa.duration("us")) + + # Tuple of bytes + a = pa.array((b"hello", b"world")) + self.assertEqual(a.type, pa.binary()) + + # Tuple of decimals + a = pa.array((Decimal("1.23"), Decimal("4.56"))) + self.assertEqual(a.type, pa.decimal128(3, 2)) + + @unittest.skipIf(not have_pandas, pandas_requirement_message) + def test_pandas_series(self): + """Test type inference from pandas Series for all types.""" + import pyarrow as pa + import pandas as pd + + # ========== numpy-backed Series (default) ========== + + # Integer Series (numpy int64) + s = pd.Series([1, 2, 3]) + a = pa.array(s) + self.assertEqual(a.type, pa.int64()) + + # Float Series (numpy float64) + s = pd.Series([1.0, 2.0, 3.0]) + a = pa.array(s) + self.assertEqual(a.type, pa.float64()) + + # String Series (object dtype) + s = pd.Series(["a", "b", "c"]) + a = pa.array(s) + self.assertEqual(a.type, pa.string()) + + # Boolean Series (numpy bool) + s = pd.Series([True, False, True]) + a = pa.array(s) + self.assertEqual(a.type, pa.bool_()) + + # Series with None (converts to float due to numpy) + s = pd.Series([1, 2, None]) + a = pa.array(s) + self.assertEqual(a.type, pa.float64()) + + # Date Series (object dtype with date objects) + s = pd.Series([datetime.date(2024, 1, 1), datetime.date(2024, 1, 2)]) + a = pa.array(s) + self.assertEqual(a.type, pa.date32()) + + # Datetime Series (pandas Timestamp, numpy datetime64[ns]) + s = pd.Series(pd.to_datetime(["2024-01-01", "2024-01-02"])) Review Comment: lets also test datetime.datetime and pd.Timestamp with timezone -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
