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

Fix approximate percentile reduction UnsupportedOperationException #5160

Merged
merged 1 commit into from
Apr 8, 2022
Merged

Fix approximate percentile reduction UnsupportedOperationException #5160

Show file tree
Hide file tree
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
16 changes: 16 additions & 0 deletions integration_tests/src/main/python/hash_aggregate_test.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1279,6 +1279,22 @@ def test_hash_groupby_approx_percentile_reduction(aqe_enabled):
lambda spark: gen_df(spark, [('v', DoubleGen())], length=100),
[0.05, 0.25, 0.5, 0.75, 0.95], conf, reduction = True)

@incompat
@pytest.mark.parametrize('aqe_enabled', ['false', 'true'], ids=idfn)
def test_hash_groupby_approx_percentile_reduction_single_row(aqe_enabled):
conf = {'spark.sql.adaptive.enabled': aqe_enabled}
compare_percentile_approx(
lambda spark: gen_df(spark, [('v', DoubleGen())], length=1),
[0.05, 0.25, 0.5, 0.75, 0.95], conf, reduction = True)

@incompat
@pytest.mark.parametrize('aqe_enabled', ['false', 'true'], ids=idfn)
def test_hash_groupby_approx_percentile_reduction_no_rows(aqe_enabled):
conf = {'spark.sql.adaptive.enabled': aqe_enabled}
compare_percentile_approx(
lambda spark: gen_df(spark, [('v', DoubleGen())], length=0),
[0.05, 0.25, 0.5, 0.75, 0.95], conf, reduction = True)

@incompat
@pytest.mark.parametrize('aqe_enabled', ['false', 'true'], ids=idfn)
def test_hash_groupby_approx_percentile_byte(aqe_enabled):
Expand Down
10 changes: 7 additions & 3 deletions sql-plugin/src/main/scala/com/nvidia/spark/rapids/GpuApproximatePercentile.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -21,6 +21,7 @@ import ai.rapids.cudf.{DType, GroupByAggregation, ReductionAggregation}
import com.nvidia.spark.rapids.GpuCast.doCast
import com.nvidia.spark.rapids.shims.ShimExpression

import org.apache.spark.sql.catalyst.InternalRow
import org.apache.spark.sql.catalyst.expressions.{AttributeReference, Expression}
import org.apache.spark.sql.catalyst.expressions.aggregate.ApproximatePercentile
import org.apache.spark.sql.catalyst.util.ArrayData
Expand Down Expand Up @@ -64,9 +65,12 @@ case class GpuApproximatePercentile (
// Attributes of fields in the aggregation buffer.
override def aggBufferAttributes: Seq[AttributeReference] = outputBuf :: Nil

// initialValues is only used in reduction and this is not currently supported
override lazy val initialValues: Seq[GpuExpression] = throw new UnsupportedOperationException(
"approx_percentile does not support reduction")
override lazy val initialValues: Seq[GpuLiteral] = Seq(GpuLiteral(
InternalRow(
ArrayData.toArrayData(Array.empty), // centroids (mean, weight)
0d, // min
0d), // max
CudfTDigest.dataType))

// the update expression will create a t-digest (List[Struct[Double, Double])
override lazy val updateAggregates: Seq[CudfAggregate] =
Expand Down