asddfl created SPARK-60648:
------------------------------
Summary: pandas-on-Spark comparison between a boolean Series and a
float Series raises a SQL type error
Key: SPARK-60648
URL: https://issues.apache.org/jira/browse/SPARK-60648
Project: Spark
Issue Type: Bug
Components: Pandas API on Spark, PySpark
Affects Versions: 4.2.0
Environment: Platform: Ubuntu 24.04 Linux-6.17.0-29-generic x86_64
Python: 3.11.17
pyspark 4.2.0
findspark 2.0.1
pandas 3.0.6
numpy 2.4.6
Reporter: asddfl
Comparing a boolean Series derived from c0 != 0 with the original float Series
via .gt() succeeds in Pandas but fails in the pandas API on Spark. The
generated SQL compares BOOLEAN and DOUBLE operands without compatible numeric
promotion.
{code:python}
import pandas as pd
from pyspark.sql import SparkSession
import pyspark.pandas as ps
t0 = pd.DataFrame({'c0': [1.0]})
print("Pandas:", flush=True)
result = (t0['c0'] != 0).gt(t0['c0'])
print(result, flush=True)
spark = (
SparkSession.builder
.master("local[2]")
.config("spark.sql.ansi.enabled", "false")
.config("spark.sql.shuffle.partitions", "2")
.getOrCreate()
)
t0 = ps.DataFrame({'c0': [1.0]})
print("PySpark Pandas:", flush=True)
result = (t0['c0'] != 0).gt(t0['c0'])
print(result, flush=True)
spark.stop()
{code}
{code:bash}
Pandas:
0 False
Name: c0, dtype: bool
PySpark Pandas:
Traceback (most recent call last):
File "/test.py", line 24, in <module>
result = (t0['c0'] != 0).gt(t0['c0'])
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/opt/conda/envs/lib/python3.11/site-packages/pyspark/pandas/series.py",
line 815, in gt
return self > other
^^^^^^^^^^^^
File "/opt/conda/envs/lib/python3.11/site-packages/pyspark/pandas/base.py",
line 507, in __gt__
return self._dtype_op.gt(self, other)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File
"/opt/conda/envs/lib/python3.11/site-packages/pyspark/pandas/data_type_ops/boolean_ops.py",
line 395, in gt
return column_op(PySparkColumn.__gt__)(left, right)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/opt/conda/envs/lib/python3.11/site-packages/pyspark/pandas/base.py",
line 246, in wrapper
self._internal.spark_frame.select(scol).schema[0],
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File
"/opt/conda/envs/lib/python3.11/site-packages/pyspark/sql/classic/dataframe.py",
line 1101, in select
jdf = self._jdf.select(self._jcols(*cols))
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/opt/conda/envs/lib/python3.11/site-packages/py4j/java_gateway.py",
line 1362, in __call__
return_value = get_return_value(
^^^^^^^^^^^^^^^^^
File
"/opt/conda/envs/lib/python3.11/site-packages/pyspark/errors/exceptions/captured.py",
line 257, in deco
raise converted from None
pyspark.errors.exceptions.captured.AnalysisException:
[DATATYPE_MISMATCH.BINARY_OP_DIFF_TYPES] Cannot resolve "(CASE WHEN ((NOT (c0 =
0)) IS NULL) THEN true ELSE (NOT (c0 = 0)) END > c0)" due to data type
mismatch: the left and right operands of the binary operator have incompatible
types ("BOOLEAN" and "DOUBLE"). SQLSTATE: 42K09;
'Project [unresolvedalias((CASE WHEN isnull(NOT (c0#1 = cast(0 as double)))
THEN true ELSE NOT (c0#1 = cast(0 as double)) END > c0#1))]
+- Project [__index_level_0__#0L, c0#1, monotonically_increasing_id() AS
__natural_order__#2L]
+- LocalRelation [__index_level_0__#0L, c0#1]
{code}
*Expected behavior*
For c0 = 1.0, return a boolean Series containing False: the comparison is True
> 1.0, which is false in Pandas.
*Actual behavior*
The call raises AnalysisException with condition
DATATYPE_MISMATCH.BINARY_OP_DIFF_TYPES and SQLSTATE 42K09. The error identifies
BOOLEAN and DOUBLE as incompatible operand types.
--
This message was sent by Atlassian Jira
(v10.3.26#10030026)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
