Max Gekk created SPARK-57462:
--------------------------------
Summary: Add PySpark support for nanosecond-precision timestamp
types
Key: SPARK-57462
URL: https://issues.apache.org/jira/browse/SPARK-57462
Project: Spark
Issue Type: Sub-task
Components: PySpark
Affects Versions: 5.0.0
Reporter: Max Gekk
Umbrella: SPARK-56822 (Timestamps with nanosecond precision).
Expose the nanosecond-capable timestamp types TIMESTAMP_NTZ(p) and
TIMESTAMP_LTZ(p) (p in 7-9) in PySpark, to reach parity with the microsecond
TimestampType / TimestampNTZType. Today python/pyspark/sql/types.py defines
only the microsecond singletons; there are no nanosecond type classes anywhere
under python/.
Scope:
- Add TimestampNTZNanosType and TimestampLTZNanosType (parameterized by
precision) to python/pyspark/sql/types.py, including toInternal / fromInternal
handling for the (epoch micros, nanos-within-micro) representation and JSON/DDL
(de)serialization.
- Mirror in the Spark Connect Python path (python/pyspark/sql/connect/types.py)
and the generated proto stubs (types_pb2.pyi) - depends on the Connect proto
work (SPARK-57160/57161).
- Decide type inference for datetime.datetime (keep inferring microsecond
TimestampType by default; nanos only via an explicit schema).
- Arrow / pandas conversion depends on the Arrow mapping (SPARK-57159).
- Tests: python/pyspark/sql/tests for createDataFrame / collect roundtrip with
explicit nanos schema, DDL parsing, and equality/repr.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
