[SPARK-36031] Keep same behavior with pandas for operations of series with nan - ASF JIRA

XML

Word

Printable

JSON

There are many operations for series containing nan values that don't follow pandas' behavior. We ought to adjust that. Please refer to sub-tasks.

relates to

SPARK-34849 SPIP: Support pandas API layer on PySpark

1.	Support arithmetic operations of Series containing Decimal(np.nan)	Resolved	Unassigned
2.	Adjust exponentiation between Series with missing values and bool literal to follow pandas	Resolved	Yikun Jiang
3.	Adjust `astype` of fractional Series with missing values to follow pandas	Resolved	Xinrong Meng