beliefer commented on a change in pull request #29800: URL: https://github.com/apache/spark/pull/29800#discussion_r513138536
########## File path: sql/core/src/main/scala/org/apache/spark/sql/execution/window/WindowFunctionFrame.scala ########## @@ -151,10 +173,93 @@ final class OffsetWindowFunctionFrame( } inputIndex += 1 } +} - override def currentLowerBound(): Int = throw new UnsupportedOperationException() +/** + * The unbounded offset window frame is an internal window frame just used to optimize the + * performance for the window function that returns the value of the input column offset + * by a number of rows within the partition and has specified ROWS BETWEEN UNBOUNDED PRECEDING + * AND UNBOUNDED FOLLOWING. The internal window frame is not a popular window frame cannot be + * specified and used directly by the users. + * The unbounded offset window frame calculates frames containing NTH_VALUE statements. + * The unbounded offset window frame return the same value for all rows in the window partition. + */ +class UnboundedOffsetWindowFunctionFrame( + target: InternalRow, + ordinal: Int, + expressions: Array[OffsetWindowSpec], + inputSchema: Seq[Attribute], + newMutableProjection: (Seq[Expression], Seq[Attribute]) => MutableProjection, + offset: Int) + extends OffsetWindowFunctionFrameBase( + target, ordinal, expressions, inputSchema, newMutableProjection, offset) { - override def currentUpperBound(): Int = throw new UnsupportedOperationException() + override def prepare(rows: ExternalAppendOnlyUnsafeRowArray): Unit = { + input = rows + if (offset > input.length) { + fillDefaultValue(EmptyRow) + } else { + inputIterator = input.generateIterator() + // drain the first few rows if offset is larger than one + inputIndex = 0 + while (inputIndex < offset - 1) { + if (inputIterator.hasNext) inputIterator.next() + inputIndex += 1 + } + val r = WindowFunctionFrame.getNextOrNull(inputIterator) + projection(r) + } + } + + override def write(index: Int, current: InternalRow): Unit = { + // The results are the same for each row in the partition, and have been evaluated in prepare. + // Don't need to recalculate here. + } +} + +/** + * The unbounded preceding offset window frame is an internal window frame just used to optimize + * the performance for the window function that returns the value of the input column offset + * by a number of rows within the partition and has specified ROWS BETWEEN UNBOUNDED PRECEDING + * AND CURRENT ROW. The internal window frame is not a popular window frame cannot be specified + * and used directly by the users. + * The unbounded preceding offset window frame calculates frames containing NTH_VALUE statements. + * The unbounded preceding offset window frame return the same value for rows which index + * (starting from 1) equal to or greater than offset in the window partition. + */ +class UnboundedPrecedingOffsetWindowFunctionFrame( + target: InternalRow, + ordinal: Int, + expressions: Array[OffsetWindowSpec], + inputSchema: Seq[Attribute], + newMutableProjection: (Seq[Expression], Seq[Attribute]) => MutableProjection, + offset: Int) + extends OffsetWindowFunctionFrameBase( + target, ordinal, expressions, inputSchema, newMutableProjection, offset) { + + var selectedRow: UnsafeRow = null + + override def prepare(rows: ExternalAppendOnlyUnsafeRowArray): Unit = { + input = rows + inputIterator = input.generateIterator() + // drain the first few rows if offset is larger than one + inputIndex = 0 + while (inputIndex < offset - 1) { + if (inputIterator.hasNext) inputIterator.next() + inputIndex += 1 + } + if (inputIndex >= 0 && inputIndex < input.length) { Review comment: Yes. ---------------------------------------------------------------- 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