zhengruifeng commented on code in PR #53721: URL: https://github.com/apache/spark/pull/53721#discussion_r2670765706
########## python/pyspark/tests/upstream/pyarrow/test_pyarrow_type_coercion.py: ########## @@ -0,0 +1,1187 @@ +# +# 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 math +import unittest + +from pyspark.testing.utils import ( + have_pandas, + have_pyarrow, + pandas_requirement_message, + pyarrow_requirement_message, +) + + +# Test pa.array type coercion behavior when creating arrays with explicit type parameter. +# This test monitors the behavior of PyArrow's type coercion to ensure PySpark's assumptions +# about PyArrow behavior remain valid across versions. +# +# Key findings: +# 1. Numeric coercion (int <-> float, int size narrowing/widening) works via safe casting +# 2. String coercion (int -> string) requires explicit pc.cast(), not implicit via type param +# 3. Decimal coercion (int -> decimal) works directly +# 4. Boolean coercion (int -> bool) does NOT work implicitly, requires explicit casting +# +# Test coverage includes: +# - Missing values (None, NaN, NaT) +# - Empty datasets +# - Invalid values and error handling +# - All Spark/PyArrow/Pandas datatypes +# - Python, Pandas, and NumPy input types +# - PySpark pandas_options for to_pandas() + + [email protected](not have_pyarrow, pyarrow_requirement_message) +class PyArrowTypeCoercionTests(unittest.TestCase): + """Test PyArrow's type coercion behavior for pa.array with explicit type parameter.""" + + # ============================================================ + # SECTION 1: Missing Values Tests (None, NaN, NaT) + # ============================================================ + + def test_none_values_in_coercion(self): + """Test that None values are preserved during type coercion.""" + import pyarrow as pa + import pyarrow.compute as pc + + # int with None -> float64 + a = pa.array([1, None, 3], type=pa.float64()) + self.assertEqual(a.type, pa.float64()) + self.assertEqual(a.to_pylist(), [1.0, None, 3.0]) + self.assertEqual(a.null_count, 1) + + # int with None -> decimal128 + a = pa.array([1, None, 3], type=pa.decimal128(10, 0)) + self.assertEqual(a.type, pa.decimal128(10, 0)) + self.assertEqual(a.to_pylist(), [Decimal("1"), None, Decimal("3")]) + self.assertEqual(a.null_count, 1) + + # int with None -> int32 (narrowing) + a = pa.array([1, None, 3], type=pa.int32()) + self.assertEqual(a.type, pa.int32()) + self.assertEqual(a.to_pylist(), [1, None, 3]) + self.assertEqual(a.null_count, 1) + + # explicit cast to string with None + int_arr = pa.array([1, None, 3]) + str_arr = pc.cast(int_arr, pa.string()) Review Comment: let's test `cast` in separate PRs -- 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]
