singhpraveen2010 opened a new pull request, #54628: URL: https://github.com/apache/spark/pull/54628
### What changes were proposed in this pull request? This resolves the open issue: https://issues.apache.org/jira/browse/SPARK-54665 (https://issues.apache.org/jira/browse/SPARK-54665) This PR resolves an inconsistency in pandas-on-Spark (`pyspark.pandas`) where comparing a boolean `Series` to a string literal returns `True` for equality (`==`) and `False` for inequality (`!=`). This PR explicitly overrides the `eq` and `ne` methods in `BooleanOps` (`python/pyspark/pandas/data_type_ops/boolean_ops.py`) to correctly handle string comparisons by returning a boolean series filled with `False` (for `eq`) and `True` (for `ne`), aligning the behavior with native pandas semantics. ### Why are the changes needed? This change is needed to maintain strict API and behavioral parity with native pandas. Currently, in `pyspark.pandas` 4.0.1: `ps.Series([True, False]) == 'True'` incorrectly yields `[True, True]`. In native `pandas`: `pd.Series([True, False]) == 'True'` correctly yields `[False, False]`. This divergence causes silent logical errors when migrating existing pandas workloads to Spark. By fixing this, we ensure predictable and correct comparison logic for end-users. ### Does this PR introduce _any_ user-facing change? Yes, this fixes a bug in behavioral semantics. **Previous Behavior (pandas-on-Spark):** ```python import pyspark.pandas as ps s = ps.Series([True, False]) s == "True" # 0 True # 1 True ``` **New Behavior (Matching native pandas):** ```python import pyspark.pandas as ps s = ps.Series([True, False]) s == "True" # 0 False # 1 False ``` ### How was this patch tested? This patch was tested by adding explicit unit tests verifying equality (`==`) and inequality (`!=`) comparisons between boolean series and string literals. Added tests in `python/pyspark/pandas/tests/data_type_ops/test_boolean_ops.py`: - `BooleanOpsTests.test_eq_ne_with_string` (indirectly via `test_eq` and `test_ne` updates) - `BooleanExtensionOpsTestsMixin` updates to ensure extension types are also covered. Tests can be run locally via: ```bash ./python/run-tests --modules pyspark-pandas-slow --testnames "pyspark.pandas.tests.data_type_ops.test_boolean_ops" ``` ### Was this patch authored or co-authored using generative AI tooling? co-authored using: Gemini CLI (Model: Gemini 2.5 Pro) -- 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]
