cloud-fan commented on code in PR #53370:
URL: https://github.com/apache/spark/pull/53370#discussion_r2596824375
##########
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/expressions/timeExpressions.scala:
##########
@@ -770,6 +770,22 @@ abstract class IntegralToTimeBase
val nanos = Math.multiplyExact(input.asInstanceOf[Number].longValue(),
upScaleFactor)
validateTimeNanos(nanos)
}
+
+ override protected def doGenCode(ctx: CodegenContext, ev: ExprCode):
ExprCode = {
+ nullSafeCodeGen(ctx, ev, c => {
+ val nanos = if (upScaleFactor == 1) {
+ c
+ } else {
+ s"java.lang.Math.multiplyExact($c, ${upScaleFactor}L)"
+ }
+ s"""
+ |${ev.value} = $nanos;
+ |if (${ev.value} < 0L || ${ev.value} >= ${NANOS_PER_DAY}L) {
+ | ${ev.isNull} = true;
Review Comment:
I just realized that we are following non-ANSI behavior here: return null
instead of failing for invalid input. Why is this decision? Do other databases
do the same?
--
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]
