Github user viirya commented on a diff in the pull request:
https://github.com/apache/spark/pull/21231#discussion_r186925349
--- Diff:
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/expressions/SortOrder.scala
---
@@ -147,7 +148,40 @@ case class SortPrefix(child: SortOrder) extends
UnaryExpression {
(!child.isAscending && child.nullOrdering == NullsLast)
}
- override def eval(input: InternalRow): Any = throw new
UnsupportedOperationException
+ private lazy val calcPrefix: Any => Long = child.child.dataType match {
+ case BooleanType => (raw) =>
+ if (raw.asInstanceOf[Boolean]) 1 else 0
+ case DateType | TimestampType | _: IntegralType => (raw) =>
+ raw.asInstanceOf[java.lang.Number].longValue()
+ case FloatType | DoubleType => (raw) => {
+ val dVal = raw.asInstanceOf[java.lang.Number].doubleValue()
+ DoublePrefixComparator.computePrefix(dVal)
+ }
+ case StringType => (raw) =>
+ StringPrefixComparator.computePrefix(raw.asInstanceOf[UTF8String])
+ case BinaryType => (raw) =>
+ BinaryPrefixComparator.computePrefix(raw.asInstanceOf[Array[Byte]])
+ case dt: DecimalType if dt.precision <= Decimal.MAX_LONG_DIGITS =>
+ _.asInstanceOf[Decimal].toUnscaledLong
+ case dt: DecimalType if dt.precision - dt.scale <=
Decimal.MAX_LONG_DIGITS => (raw) => {
+ val value = raw.asInstanceOf[Decimal]
+ val p = Decimal.MAX_LONG_DIGITS
+ val s = p - (dt.precision - dt.scale)
--- End diff --
This is a nit but we can move L168 & L169 out of closure too. Like:
```scala
case dt: DecimalType if dt.precision - dt.scale <= Decimal.MAX_LONG_DIGITS
=>
val p = Decimal.MAX_LONG_DIGITS
val s = p - (dt.precision - dt.scale)
(raw) => {
val value = raw.asInstanceOf[Decimal]
if (value.changePrecision(p, s)) value.toUnscaledLong else Long.MinValue
}
```
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
