Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Use merge sort for struct types in non-key columns #4459

Merged
merged 4 commits into from
Jan 11, 2022
Merged

Use merge sort for struct types in non-key columns #4459

Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
d0988f3
Support struct type in sort merge
firestarman Dec 21, 2021
b23c568
comment update
firestarman Jan 5, 2022
e3a5ca3
Fix the test error
firestarman Jan 6, 2022
398c674
Address the comments
firestarman Jan 10, 2022
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
30 changes: 20 additions & 10 deletions sql-plugin/src/main/scala/com/nvidia/spark/rapids/SortUtils.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,5 +1,5 @@
/*
* Copyright (c) 2019-2021, NVIDIA CORPORATION.
* Copyright (c) 2019-2022, NVIDIA CORPORATION.
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
Expand All @@ -22,7 +22,8 @@ import scala.collection.mutable.ArrayBuffer
import ai.rapids.cudf.{ColumnVector, NvtxColor, OrderByArg, Table}

import org.apache.spark.sql.catalyst.expressions.{Alias, Attribute, AttributeReference, BoundReference, Expression, NullsFirst, NullsLast, SortOrder}
import org.apache.spark.sql.types.DataType
import org.apache.spark.sql.rapids.execution.TrampolineUtil
import org.apache.spark.sql.types.{ArrayType, DataType, MapType}
import org.apache.spark.sql.vectorized.ColumnarBatch

object SortUtils extends Arm {
Expand Down Expand Up @@ -210,6 +211,20 @@ class GpuSorter(
}
}

private[this] lazy val hasNestedInKeyColumns = cpuOrderingInternal.exists ( order =>
DataTypeUtils.isNestedType(order.child.dataType)
)

/** (This can be removed once https://github.com/rapidsai/cudf/issues/8050 is addressed) */
private[this] lazy val hasUnsupportedNestedInRideColumns = {
val keyColumnIndices = cpuOrderingInternal.map(_.child.asInstanceOf[BoundReference].ordinal)
val rideColumnIndices = projectedBatchTypes.indices.toSet -- keyColumnIndices
rideColumnIndices.exists { idx =>
TrampolineUtil.dataTypeExistsRecursively(projectedBatchTypes(idx),
t => t.isInstanceOf[ArrayType] || t.isInstanceOf[MapType])
}
}

/**
* Merge multiple batches together. All of these batches should be the output of
* `appendProjectedColumns` and the output of this will also be in that same format.
Expand All @@ -226,15 +241,10 @@ class GpuSorter(
batches.foreach { cb =>
tabs += GpuColumnVector.from(cb)
}
// In the current version of cudf merge does not work for structs or lists (nested types)
// In the current version of cudf merge does not work for lists and maps.
// This should be fixed by https://github.com/rapidsai/cudf/issues/8050
val hasNested = {
val tab = tabs.head
(0 until tab.getNumberOfColumns).exists { i =>
tab.getColumn(i).getType.isNestedType
}
}
if (hasNested) {
// Nested types in sort key columns is not supported either.
if (hasNestedInKeyColumns || hasUnsupportedNestedInRideColumns) {
// so as a work around we concatenate all of the data together and then sort it.
// It is slower, but it works
withResource(Table.concatenate(tabs: _*)) { concatenated =>
Expand Down