Skip to content

[SPARK-24999][SQL]Reduce unnecessary 'new' memory operations #21968

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

Closed
heary-cao wants to merge 1 commit into from
Closed

[SPARK-24999][SQL]Reduce unnecessary 'new' memory operations #21968

Changes from all commits
Commits
Show all changes
1 commit
Select commit
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: 18 additions & 12 deletions ...re/src/main/scala/org/apache/spark/sql/execution/aggregate/RowBasedHashMapGenerator.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -48,6 +48,12 @@ class RowBasedHashMapGenerator(
val keySchema = ctx.addReferenceObj("keySchemaTerm", groupingKeySchema)
val valueSchema = ctx.addReferenceObj("valueSchemaTerm", bufferSchema)

val numVarLenFields = groupingKeys.map(_.dataType).count {
Copy link
Contributor

@cloud-fan cloud-fan Aug 31, 2018

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

groupingKeys.map(_.dataType).count(dt => !UnsafeRow.isFixedLength(dt))

Copy link
Contributor Author

@heary-cao heary-cao Aug 31, 2018

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

ok, thanks.

case dt if UnsafeRow.isFixedLength(dt) => false
// TODO: consider large decimal and interval type
case _ => true
}

s"""
| private org.apache.spark.sql.catalyst.expressions.RowBasedKeyValueBatch batch;
| private int[] buckets;
Expand All @@ -60,6 +66,7 @@ class RowBasedHashMapGenerator(
| private long emptyVOff;
| private int emptyVLen;
| private boolean isBatchFull = false;
| private org.apache.spark.sql.catalyst.expressions.codegen.UnsafeRowWriter agg_rowWriter;
|
|
| public $generatedClassName(
Expand All @@ -75,6 +82,9 @@ class RowBasedHashMapGenerator(
| emptyVOff = Platform.BYTE_ARRAY_OFFSET;
| emptyVLen = emptyBuffer.length;
|
| agg_rowWriter = new org.apache.spark.sql.catalyst.expressions.codegen.UnsafeRowWriter(
| ${groupingKeySchema.length}, ${numVarLenFields * 32});
|
| buckets = new int[numBuckets];
| java.util.Arrays.fill(buckets, -1);
| }
Expand Down Expand Up @@ -112,12 +122,6 @@ class RowBasedHashMapGenerator(
*
*/
protected def generateFindOrInsert(): String = {
val numVarLenFields = groupingKeys.map(_.dataType).count {
case dt if UnsafeRow.isFixedLength(dt) => false
// TODO: consider large decimal and interval type
case _ => true
}

val createUnsafeRowForKey = groupingKeys.zipWithIndex.map { case (key: Buffer, ordinal: Int) =>
key.dataType match {
case t: DecimalType =>
Expand All @@ -130,6 +134,12 @@ class RowBasedHashMapGenerator(
}
}.mkString(";\n")

val resetNullBits = if (groupingKeySchema.map(_.nullable).forall(_ == false)) {
""
} else {
"agg_rowWriter.zeroOutNullBytes();"
}

s"""
|public org.apache.spark.sql.catalyst.expressions.UnsafeRow findOrInsert(${
groupingKeySignature}) {
Expand All @@ -140,12 +150,8 @@ class RowBasedHashMapGenerator(
| // Return bucket index if it's either an empty slot or already contains the key
| if (buckets[idx] == -1) {
| if (numRows < capacity && !isBatchFull) {
| // creating the unsafe for new entry
| org.apache.spark.sql.catalyst.expressions.codegen.UnsafeRowWriter agg_rowWriter
| = new org.apache.spark.sql.catalyst.expressions.codegen.UnsafeRowWriter(
| ${groupingKeySchema.length}, ${numVarLenFields * 32});
| agg_rowWriter.reset(); //TODO: investigate if reset or zeroout are actually needed
| agg_rowWriter.zeroOutNullBytes();
| agg_rowWriter.reset();
| $resetNullBits
| ${createUnsafeRowForKey};
| org.apache.spark.sql.catalyst.expressions.UnsafeRow agg_result
| = agg_rowWriter.getRow();
Expand Down