cloud-fan commented on a change in pull request #28673: URL: https://github.com/apache/spark/pull/28673#discussion_r432817075
########## File path: sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/util/DateTimeFormatterHelper.scala ########## @@ -118,6 +118,31 @@ trait DateTimeFormatterHelper { s"before Spark 3.0, or set to CORRECTED and treat it as an invalid datetime string.", e) } + // When legacy time parser policy set to EXCEPTION, check whether we will get different results + // between legacy formatter and new formatter. If new formatter fails but legacy formatter works, + // throw a SparkUpgradeException. On the contrary, if the legacy policy set to CORRECTED, + // DateTimeParseException will address by the caller side. + protected def checkDiffFormatResult[T <: Date]( + d: T, + legacyFormatFunc: T => String): PartialFunction[Throwable, String] = { + case e if needConvertToSparkUpgradeException(e) => + val resultCandidate = try { + legacyFormatFunc(d) + } catch { + case _: Throwable => throw e + } + throw new SparkUpgradeException("3.0", s"Fail to format '$resultCandidate' in the new" + + s" formatter. You can set ${SQLConf.LEGACY_TIME_PARSER_POLICY.key} to LEGACY to restore" + + s" the behavior before Spark 3.0, or set to CORRECTED and treat it as an invalid" + + s" datetime string.", e) + } + + private def needConvertToSparkUpgradeException(e: Throwable): Boolean = e match { + case _: DateTimeException | _: ArrayIndexOutOfBoundsException Review comment: This seems like a Java bug to me to throw `ArrayIndexOutOfBoundsException`. Shall we fail the formatter creation when there are more than 10 `y` in the pattern string? And document this limitation as well. ---------------------------------------------------------------- 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. For queries about this service, please contact Infrastructure at: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org