uros-b commented on code in PR #56616:
URL: https://github.com/apache/spark/pull/56616#discussion_r3444978176
##########
sql/catalyst/src/test/scala/org/apache/spark/sql/catalyst/expressions/DateExpressionsSuite.scala:
##########
@@ -1743,6 +1743,62 @@ class DateExpressionsSuite extends SparkFunSuite with
ExpressionEvalHelper {
}
}
+ test("SPARK-57526: timestamp_nanos builds a TIMESTAMP_LTZ(9) from
nanoseconds") {
+ import org.apache.spark.sql.catalyst.util.TimestampNanosTestUtils._
+
+ // DECIMAL input is accepted as-is; a wide DECIMAL(38, 0) holds every
input below.
+ def tsNanos(n: BigInt): NanosToTimestamp =
+ NanosToTimestamp(Literal.create(Decimal(BigDecimal(n), 38, 0),
DecimalType(38, 0)))
+
+ assert(tsNanos(0).dataType === TimestampLTZNanosType(9))
+
+ // The JIRA example: 1230219000123456789 ns -> 1230219000123456 micros +
789 ns.
+ checkEvaluation(tsNanos(BigInt("1230219000123456789")),
nanosVal(1230219000123456L, 789))
+
+ // An integral argument is accepted directly (widened to BigInteger),
exercising the
+ // IntegralType eval/codegen path rather than the DECIMAL one.
+ checkEvaluation(
+ NanosToTimestamp(Literal(1230219000123456789L)),
nanosVal(1230219000123456L, 789))
+ checkEvaluation(NanosToTimestamp(Literal(-1L)), nanosVal(-1L, 999))
+ checkEvaluation(NanosToTimestamp(Literal(1000)), nanosVal(1L, 0))
Review Comment:
Nit about integral-width coverage: the catalyst test exercises Int
(Literal(1000)) and Long, which is enough to cover the (long) $c codegen cast,
but a TINYINT/SMALLINT case would fully nail the IntegralType branch.
--
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]
