[
https://issues.apache.org/jira/browse/SPARK-57834?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Max Gekk updated SPARK-57834:
-----------------------------
Shepherd: Max Gekk
> Support nanosecond-precision timestamp endpoints in the sequence function
> -------------------------------------------------------------------------
>
> Key: SPARK-57834
> URL: https://issues.apache.org/jira/browse/SPARK-57834
> Project: Spark
> Issue Type: Sub-task
> Components: SQL
> Affects Versions: 4.3.0
> Reporter: Max Gekk
> Priority: Major
>
> This sub-task is part of the umbrella SPARK-56822 (timestamps with nanosecond
> precision).
> h2. Problem
> {{Sequence}} (collectionOperations.scala ~L3322-3378) allows only
> {{TimestampType | TimestampNTZType}} endpoints and steps through {{Long}}
> micros, so nanosecond start/stop raise an analysis error.
> h2. Goal
> Generate an array of nanosecond timestamps from nanosecond start/stop with an
> interval step, preserving precision/family and the {{nanosWithinMicro}}
> remainder across steps.
> h2. Scope
> Extend {{Sequence.checkInputDataTypes}} and the implementation to accept
> {{AnyTimestampNanoType}}; add a {{TimestampNanosVal}} stepping path (day-time
> interval step, consistent with the SPARK-57501 arithmetic).
> h2. Acceptance criteria
> * {{sequence(ts_nanos_start, ts_nanos_stop, interval)}} yields
> correctly-spaced nanosecond timestamps of the input type.
> h2. Testing
> {{CollectionExpressionsSuite}}; nanos golden files.
> h2. Dependencies
> None hard - steps via resolved SPARK-57501 arithmetic; coordinate with the
> overflow/range-hardening sub-task for step overflow.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
