cloud-fan commented on a change in pull request #26473: [WIP] Strict parsing of
day-time strings to intervals
URL: https://github.com/apache/spark/pull/26473#discussion_r345338799
##########
File path:
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/util/IntervalUtils.scala
##########
@@ -250,6 +262,66 @@ object IntervalUtils {
}
}
+ private val signRe = "(?<sign>[+|-])"
+ private val dayRe = "(?<day>\\d+)"
+ private val hourRe = "(?<hour>\\d{1,2})"
+ private val minuteRe = "(?<minute>\\d{1,2})"
+ private val secondRe = "(?<second>(\\d{1,2})(\\.(\\d{1,9}))?)"
+
+ private val dayTimePattern = Map(
+ (MINUTE, SECOND) -> s"^$signRe?$minuteRe:$secondRe$$".r,
+ (HOUR, MINUTE) -> s"^$signRe?$hourRe:$minuteRe$$".r,
+ (HOUR, SECOND) -> s"^$signRe?$hourRe:$minuteRe:$secondRe$$".r,
+ (DAY, HOUR) -> s"^$signRe?$dayRe $hourRe$$".r,
+ (DAY, MINUTE) -> s"^$signRe?$dayRe $hourRe:$minuteRe$$".r,
+ (DAY, SECOND) -> s"^$signRe?$dayRe $hourRe:$minuteRe:$secondRe$$".r
+ )
+
+ private def unitsRange(start: IntervalUnit, end: IntervalUnit):
Seq[IntervalUnit] = {
+ (start.id to end.id).map(IntervalUnit(_))
+ }
+
+ private def fromDayTime(
+ input: String,
+ from: IntervalUnit,
+ to: IntervalUnit): CalendarInterval = {
+ require(input != null, "Interval day-time string must be not null")
+ val regexp = dayTimePattern.get(from -> to)
+ require(regexp.isDefined, s"Cannot support (interval '$input' $from to
$to) expression")
+ val pattern = regexp.get.pattern
+ val m = pattern.matcher(input)
+ require(m.matches, s"Interval string must match day-time format of
'$pattern': $input")
+
+ try {
+ var micros: Long = 0L
+ var days: Int = 0
+
+ unitsRange(to, from).foreach {
+ case unit @ DAY =>
+ val parsed = toLongWithRange(unit, m.group(unit.toString), 0,
Int.MaxValue)
+ days = Math.toIntExact(parsed)
+ case unit @ HOUR =>
+ val parsed = toLongWithRange(unit, m.group(unit.toString), 0, 23)
+ micros = Math.addExact(micros, Math.multiplyExact(parsed,
MICROS_PER_HOUR))
Review comment:
ditto, `parsed * MICROS_PER_HOUR` can never overflow
----------------------------------------------------------------
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:
[email protected]
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
