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

Decimal support for add and subtract #1561

Merged
merged 13 commits into from
Jan 21, 2021
Merged

Decimal support for add and subtract #1561

Show file tree
Hide file tree
Changes from 12 commits
Commits
Show all changes
13 commits
Select commit Hold shift + click to select a range
f3bade9
first pass decimal binary op support
rwlee Jan 14, 2021
e60f3bb
Merge branch 'branch-0.4' into rwlee/decimal_add
rwlee Jan 14, 2021
15cee82
Clean up and remove divide
rwlee Jan 15, 2021
6f88e36
Merge branch 'branch-0.4' into rwlee/decimal_add
rwlee Jan 19, 2021
207ff81
debug code for decimal integration test
rwlee Jan 19, 2021
b1bf4bf
Merge branch 'rwlee/decimal_add' of github.com:rwlee/spark-rapids int…
revans2 Jan 19, 2021
d14dcf9
Some fixes multiplication is still broken
revans2 Jan 19, 2021
31c2202
Remove unit tests, decimal expressions are true no ops
rwlee Jan 21, 2021
a956cf3
remove decimal support for mult/div
rwlee Jan 21, 2021
86156d2
Merge remote-tracking branch 'pub/branch-0.4' into rwlee/decimal_binops
rwlee Jan 21, 2021
ffc1e1b
code cleanup
rwlee Jan 21, 2021
6592b2e
fix docs and style
rwlee Jan 21, 2021
38ccde2
code cleanup
rwlee Jan 21, 2021
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
2 changes: 2 additions & 0 deletions docs/configs.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -124,6 +124,7 @@ Name | SQL Function(s) | Description | Default Value | Notes
<a name="sql.expression.Cast"></a>spark.rapids.sql.expression.Cast|`timestamp`, `tinyint`, `binary`, `float`, `smallint`, `string`, `decimal`, `double`, `boolean`, `cast`, `date`, `int`, `bigint`|Convert a column of one type of data into another type|true|None|
<a name="sql.expression.Cbrt"></a>spark.rapids.sql.expression.Cbrt|`cbrt`|Cube root|true|None|
<a name="sql.expression.Ceil"></a>spark.rapids.sql.expression.Ceil|`ceiling`, `ceil`|Ceiling of a number|true|None|
<a name="sql.expression.CheckOverflow"></a>spark.rapids.sql.expression.CheckOverflow| |CheckOverflow after arithmetic operations between DecimalType data|true|None|
<a name="sql.expression.Coalesce"></a>spark.rapids.sql.expression.Coalesce|`coalesce`|Returns the first non-null argument if exists. Otherwise, null|true|None|
<a name="sql.expression.Concat"></a>spark.rapids.sql.expression.Concat|`concat`|String concatenate NO separator|true|None|
<a name="sql.expression.Contains"></a>spark.rapids.sql.expression.Contains| |Contains|true|None|
Expand Down Expand Up @@ -193,6 +194,7 @@ Name | SQL Function(s) | Description | Default Value | Notes
<a name="sql.expression.Or"></a>spark.rapids.sql.expression.Or|`or`|Logical OR|true|None|
<a name="sql.expression.Pmod"></a>spark.rapids.sql.expression.Pmod|`pmod`|Pmod|true|None|
<a name="sql.expression.Pow"></a>spark.rapids.sql.expression.Pow|`pow`, `power`|lhs ^ rhs|true|None|
<a name="sql.expression.PromotePrecision"></a>spark.rapids.sql.expression.PromotePrecision| |PromotePrecision before arithmetic operations between DecimalType data|true|None|
<a name="sql.expression.PythonUDF"></a>spark.rapids.sql.expression.PythonUDF| |UDF run in an external python process. Does not actually run on the GPU, but the transfer of data to/from it can be accelerated.|true|None|
<a name="sql.expression.Quarter"></a>spark.rapids.sql.expression.Quarter|`quarter`|Returns the quarter of the year for date, in the range 1 to 4|true|None|
<a name="sql.expression.Rand"></a>spark.rapids.sql.expression.Rand|`random`, `rand`|Generate a random column with i.i.d. uniformly distributed values in [0, 1)|true|None|
Expand Down
192 changes: 186 additions & 6 deletions docs/supported_ops.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1068,7 +1068,7 @@ Accelerator support is described below.
<td> </td>
<td> </td>
<td> </td>
<td><b>NS</b></td>
<td>S*</td>
<td> </td>
<td> </td>
<td><b>NS</b></td>
Expand All @@ -1089,7 +1089,7 @@ Accelerator support is described below.
<td> </td>
<td> </td>
<td> </td>
<td><b>NS</b></td>
<td>S*</td>
<td> </td>
<td> </td>
<td><b>NS</b></td>
Expand All @@ -1110,7 +1110,7 @@ Accelerator support is described below.
<td> </td>
<td> </td>
<td> </td>
<td><b>NS</b></td>
<td>S*</td>
<td> </td>
<td> </td>
<td><b>NS</b></td>
Expand Down Expand Up @@ -2834,6 +2834,96 @@ Accelerator support is described below.
<td> </td>
</tr>
<tr>
<td rowSpan="4">CheckOverflow</td>
<td rowSpan="4"> </td>
<td rowSpan="4">CheckOverflow after arithmetic operations between DecimalType data</td>
<td rowSpan="4">None</td>
<td rowSpan="2">project</td>
<td>input</td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td>S*</td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
</tr>
<tr>
<td>result</td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td>S*</td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
</tr>
<tr>
<td rowSpan="2">lambda</td>
<td>input</td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td><b>NS</b></td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
</tr>
<tr>
<td>result</td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td><b>NS</b></td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
</tr>
<tr>
<td rowSpan="4">Coalesce</td>
<td rowSpan="4">`coalesce`</td>
<td rowSpan="4">Returns the first non-null argument if exists. Otherwise, null</td>
Expand Down Expand Up @@ -9859,6 +9949,96 @@ Accelerator support is described below.
<td> </td>
</tr>
<tr>
<td rowSpan="4">PromotePrecision</td>
<td rowSpan="4"> </td>
<td rowSpan="4">PromotePrecision before arithmetic operations between DecimalType data</td>
<td rowSpan="4">None</td>
<td rowSpan="2">project</td>
<td>input</td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td>S*</td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
</tr>
<tr>
<td>result</td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td>S*</td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
</tr>
<tr>
<td rowSpan="2">lambda</td>
<td>input</td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td><b>NS</b></td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
</tr>
<tr>
<td>result</td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td><b>NS</b></td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
<td> </td>
</tr>
<tr>
<td rowSpan="8">PythonUDF</td>
<td rowSpan="8"> </td>
<td rowSpan="8">UDF run in an external python process. Does not actually run on the GPU, but the transfer of data to/from it can be accelerated.</td>
Expand Down Expand Up @@ -13540,7 +13720,7 @@ Accelerator support is described below.
<td> </td>
<td> </td>
<td> </td>
<td><b>NS</b></td>
<td>S*</td>
<td> </td>
<td> </td>
<td><b>NS</b></td>
Expand All @@ -13561,7 +13741,7 @@ Accelerator support is described below.
<td> </td>
<td> </td>
<td> </td>
<td><b>NS</b></td>
<td>S*</td>
<td> </td>
<td> </td>
<td><b>NS</b></td>
Expand All @@ -13582,7 +13762,7 @@ Accelerator support is described below.
<td> </td>
<td> </td>
<td> </td>
<td><b>NS</b></td>
<td>S*</td>
<td> </td>
<td> </td>
<td><b>NS</b></td>
Expand Down
13 changes: 9 additions & 4 deletions integration_tests/src/main/python/arithmetic_ops_test.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -21,7 +21,10 @@
from spark_session import with_spark_session, is_before_spark_310
import pyspark.sql.functions as f

@pytest.mark.parametrize('data_gen', numeric_gens, ids=idfn)
decimal_gens_not_max_prec = [decimal_gen_neg_scale, decimal_gen_scale_precision,
decimal_gen_same_scale_precision, decimal_gen_64bit]

@pytest.mark.parametrize('data_gen', numeric_gens + decimal_gens_not_max_prec, ids=idfn)
def test_addition(data_gen):
data_type = data_gen.data_type
assert_gpu_and_cpu_are_equal_collect(
Expand All @@ -30,9 +33,10 @@ def test_addition(data_gen):
f.lit(-12).cast(data_type) + f.col('b'),
f.lit(None).cast(data_type) + f.col('a'),
f.col('b') + f.lit(None).cast(data_type),
f.col('a') + f.col('b')))
f.col('a') + f.col('b')),
conf=allow_negative_scale_of_decimal_conf)

@pytest.mark.parametrize('data_gen', numeric_gens, ids=idfn)
@pytest.mark.parametrize('data_gen', numeric_gens + decimal_gens_not_max_prec, ids=idfn)
def test_subtraction(data_gen):
data_type = data_gen.data_type
assert_gpu_and_cpu_are_equal_collect(
Expand All @@ -41,7 +45,8 @@ def test_subtraction(data_gen):
f.lit(-12).cast(data_type) - f.col('b'),
f.lit(None).cast(data_type) - f.col('a'),
f.col('b') - f.lit(None).cast(data_type),
f.col('a') - f.col('b')))
f.col('a') - f.col('b')),
conf=allow_negative_scale_of_decimal_conf)

@pytest.mark.parametrize('data_gen', numeric_gens, ids=idfn)
def test_multiplication(data_gen):
Expand Down
21 changes: 14 additions & 7 deletions sql-plugin/src/main/scala/com/nvidia/spark/rapids/GpuOverrides.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -41,7 +41,6 @@ import org.apache.spark.sql.execution.datasources.parquet.ParquetFileFormat
import org.apache.spark.sql.execution.datasources.text.TextFileFormat
import org.apache.spark.sql.execution.datasources.v2.{AlterNamespaceSetPropertiesExec, AlterTableExec, AtomicReplaceTableExec, BatchScanExec, CreateNamespaceExec, CreateTableExec, DeleteFromTableExec, DescribeNamespaceExec, DescribeTableExec, DropNamespaceExec, DropTableExec, RefreshTableExec, RenameTableExec, ReplaceTableExec, SetCatalogAndNamespaceExec, ShowCurrentNamespaceExec, ShowNamespacesExec, ShowTablePropertiesExec, ShowTablesExec}
import org.apache.spark.sql.execution.datasources.v2.csv.CSVScan
import org.apache.spark.sql.execution.datasources.v2.parquet.ParquetScan
import org.apache.spark.sql.execution.exchange.{BroadcastExchangeExec, ShuffleExchangeExec}
import org.apache.spark.sql.execution.joins._
import org.apache.spark.sql.execution.python._
Expand Down Expand Up @@ -706,6 +705,14 @@ object GpuOverrides {
}
}
}),
expr[PromotePrecision](
"PromotePrecision before arithmetic operations between DecimalType data",
ExprChecks.unaryProjectNotLambdaInputMatchesOutput(TypeSig.DECIMAL, TypeSig.DECIMAL),
(a, conf, p, r) => new PromotePrecisionExprMeta(a, conf, p, r)),
expr[CheckOverflow](
"CheckOverflow after arithmetic operations between DecimalType data",
ExprChecks.unaryProjectNotLambdaInputMatchesOutput(TypeSig.DECIMAL, TypeSig.DECIMAL),
(a, conf, p, r) => new CheckOverflowExprMeta(a, conf, p, r)),
expr[ToDegrees](
"Converts radians to degrees",
ExprChecks.mathUnary,
Expand Down Expand Up @@ -1377,19 +1384,19 @@ object GpuOverrides {
expr[Add](
"Addition",
ExprChecks.binaryProjectNotLambda(
TypeSig.integral + TypeSig.fp, TypeSig.numericAndInterval,
("lhs", TypeSig.integral + TypeSig.fp, TypeSig.numericAndInterval),
("rhs", TypeSig.integral + TypeSig.fp, TypeSig.numericAndInterval)),
TypeSig.numeric, TypeSig.numericAndInterval,
("lhs", TypeSig.numeric, TypeSig.numericAndInterval),
("rhs", TypeSig.numeric, TypeSig.numericAndInterval)),
(a, conf, p, r) => new BinaryExprMeta[Add](a, conf, p, r) {
override def convertToGpu(lhs: Expression, rhs: Expression): GpuExpression =
GpuAdd(lhs, rhs)
}),
expr[Subtract](
"Subtraction",
ExprChecks.binaryProjectNotLambda(
TypeSig.integral + TypeSig.fp, TypeSig.numericAndInterval,
("lhs", TypeSig.integral + TypeSig.fp, TypeSig.numericAndInterval),
("rhs", TypeSig.integral + TypeSig.fp, TypeSig.numericAndInterval)),
TypeSig.numeric, TypeSig.numericAndInterval,
("lhs", TypeSig.numeric, TypeSig.numericAndInterval),
("rhs", TypeSig.numeric, TypeSig.numericAndInterval)),
(a, conf, p, r) => new BinaryExprMeta[Subtract](a, conf, p, r) {
override def convertToGpu(lhs: Expression, rhs: Expression): GpuExpression =
GpuSubtract(lhs, rhs)
Expand Down
Loading