david-mollitor-db opened a new pull request, #58803:
URL: https://github.com/apache/spark/pull/58803
### What changes were proposed in this pull request?
`SparkDateTimeUtils.parseTimestampString` normalizes the parsed time zone
with:
```scala
val zoneId = tz.map(zoneName => getZoneId(zoneName.trim))
```
Because `getZoneId` is an instance method of the trait, the mapping function
`zoneName =>
getZoneId(zoneName.trim)` captures `this`. This PR replaces the `map` with a
direct pattern
match on the `Option`, which is behavior-preserving:
```scala
val zoneId = tz match {
case Some(zoneName) => Some(getZoneId(zoneName.trim))
case None => None
}
```
### Why are the changes needed?
A *capturing* lambda is not compiled to a cached singleton: the JVM
allocates a fresh closure
instance every time the lambda expression is evaluated (only non-capturing
lambdas get a
constant `CallSite`). `parseTimestampString` runs once per parsed timestamp
value — e.g. CSV
and JSON ingestion, casts to timestamp, and the date/time functions — so the
current code
allocates one closure per parsed timestamp on a hot path.
The pattern match produces the identical result with no closure: the `None`
case allocates
nothing, and the `Some` case allocates only the `Some` wrapper that `map`
already produced.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
Existing tests, which cover the time-zone parsing paths:
```
build/sbt 'catalyst/testOnly *DateTimeUtilsSuite *TimestampFormatterSuite'
```
All 124 tests pass.
### Was this patch authored or co-authored using generative AI tooling?
Generated-by: Isaac
This pull request and its description were written by Isaac.
--
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]
