cloud-fan commented on a change in pull request #30387:
URL: https://github.com/apache/spark/pull/30387#discussion_r547260892
##########
File path:
sql/core/src/main/scala/org/apache/spark/sql/execution/window/WindowFunctionFrame.scala
##########
@@ -147,31 +148,133 @@ class FrameLessOffsetWindowFunctionFrame(
expressions: Array[OffsetWindowFunction],
inputSchema: Seq[Attribute],
newMutableProjection: (Seq[Expression], Seq[Attribute]) =>
MutableProjection,
- offset: Int)
+ offset: Int,
+ ignoreNulls: Boolean = false)
extends OffsetWindowFunctionFrameBase(
target, ordinal, expressions, inputSchema, newMutableProjection, offset) {
+ /** The input expression of Lead/Lag. */
+ private lazy val inputExpression = expressions.toSeq.map(_.input).head
+
+ /** The index of input expression in the row. */
+ private lazy val idx = inputAttrs.zipWithIndex.find(_._1 ==
inputExpression).map(_._2).head
+
+ /** Holder the UnsafeRow where the input operator by function is not null. */
+ private var nextSelectedRow = EmptyRow
+
+ /**
+ * The number of UnsafeRows skipped to get the next UnsafeRow where
+ * the input operator by function is not null.
+ */
+ private var skipNonNullCount = 0
+
+ /** find the offset row whose input is not null */
+ private def findNextRowWithNonNullInput(): Unit = {
+ while (skipNonNullCount < offset && inputIndex < input.length) {
+ val r = WindowFunctionFrame.getNextOrNull(inputIterator)
+ if (!r.isNullAt(idx)) {
+ nextSelectedRow = r
+ skipNonNullCount += 1
+ }
+ inputIndex += 1
+ }
+ }
+
override def prepare(rows: ExternalAppendOnlyUnsafeRowArray): Unit = {
input = rows
inputIterator = input.generateIterator()
// drain the first few rows if offset is larger than zero
inputIndex = 0
- while (inputIndex < offset) {
- if (inputIterator.hasNext) inputIterator.next()
- inputIndex += 1
+ if (ignoreNulls) {
+ findNextRowWithNonNullInput
Review comment:
nit: `findNextRowWithNonNullInput()` as it has side-effect.
----------------------------------------------------------------
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]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
