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

[SPARK-22384][SQL][followup] Refine partition pruning when attribute is wrapped in Cast #21712

Closed
wants to merge 2 commits into from
Closed

[SPARK-22384][SQL][followup] Refine partition pruning when attribute is wrapped in Cast #21712

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
7b316ff
only support cast in partition predicate when it's safe
cloud-fan Jul 4, 2018
f7699ac
fix typo
cloud-fan Jul 4, 2018
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
20 changes: 20 additions & 0 deletions sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/expressions/Cast.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -134,6 +134,26 @@ object Cast {
toPrecedence > 0 && fromPrecedence > toPrecedence
}

/**
* Returns true iff we can safely cast the `from` type to `to` type without any truncating or
* precision lose, e.g. int -> long, date -> timestamp.
*/
def canSafeCast(from: AtomicType, to: AtomicType): Boolean = (from, to) match {
Copy link
Member

Choose a reason for hiding this comment

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

Is this only applied for AtomicType?

Copy link
Contributor Author

Choose a reason for hiding this comment

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

yes, at least for now. Hive partition predicate doesn't support complex type.

case _ if from == to => true
case (from: NumericType, to: DecimalType) if to.isWiderThan(from) => true
case (from: DecimalType, to: NumericType) if from.isTighterThan(to) => true
case (from, to) if legalNumericPrecedence(from, to) => true
case (DateType, TimestampType) => true
case (_, StringType) => true
case _ => false
}

private def legalNumericPrecedence(from: DataType, to: DataType): Boolean = {
val fromPrecedence = TypeCoercion.numericPrecedence.indexOf(from)
val toPrecedence = TypeCoercion.numericPrecedence.indexOf(to)
fromPrecedence >= 0 && fromPrecedence < toPrecedence
}

def forceNullable(from: DataType, to: DataType): Boolean = (from, to) match {
case (NullType, _) => true
case (_, _) if from == to => false
Expand Down
5 changes: 3 additions & 2 deletions sql/hive/src/main/scala/org/apache/spark/sql/hive/client/HiveShim.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -45,7 +45,7 @@ import org.apache.spark.sql.catalyst.analysis.NoSuchPermanentFunctionException
import org.apache.spark.sql.catalyst.catalog.{CatalogFunction, CatalogTablePartition, CatalogUtils, FunctionResource, FunctionResourceType}
import org.apache.spark.sql.catalyst.expressions._
import org.apache.spark.sql.internal.SQLConf
import org.apache.spark.sql.types.{IntegralType, StringType}
import org.apache.spark.sql.types.{AtomicType, IntegralType, StringType}
import org.apache.spark.unsafe.types.UTF8String
import org.apache.spark.util.Utils

Expand Down Expand Up @@ -660,7 +660,8 @@ private[client] class Shim_v0_13 extends Shim_v0_12 {
def unapply(expr: Expression): Option[Attribute] = {
expr match {
case attr: Attribute => Some(attr)
case Cast(child, dt, _) if !Cast.mayTruncate(child.dataType, dt) => unapply(child)
case Cast(child @ AtomicType(), dt: AtomicType, _)
if Cast.canSafeCast(child.dataType.asInstanceOf[AtomicType], dt) => unapply(child)
case _ => None
}
}
Expand Down
20 changes: 18 additions & 2 deletions sql/hive/src/test/scala/org/apache/spark/sql/hive/client/HiveClientSuite.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -24,7 +24,7 @@ import org.scalatest.BeforeAndAfterAll
import org.apache.spark.sql.catalyst.catalog._
import org.apache.spark.sql.catalyst.dsl.expressions._
import org.apache.spark.sql.catalyst.expressions._
import org.apache.spark.sql.types.LongType
import org.apache.spark.sql.types.{BooleanType, IntegerType, LongType}

// TODO: Refactor this to `HivePartitionFilteringSuite`
class HiveClientSuite(version: String)
Expand Down Expand Up @@ -122,6 +122,22 @@ class HiveClientSuite(version: String)
"aa" :: Nil)
}

test("getPartitionsByFilter: cast(chunk as int)=1 (not a valid partition predicate)") {
Copy link
Member

Choose a reason for hiding this comment

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

I ran this test with current master branch, seems it can also pass without this change?

Copy link
Contributor Author

Choose a reason for hiding this comment

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

yes this can pass, we can tell it by #12586

Cast.mayTruncate returns true for string to int, but not string to boolean. Here I just wanna test both of the cases, to improve test coverage.

testMetastorePartitionFiltering(
attr("chunk").cast(IntegerType) === 1,
20170101 to 20170103,
0 to 23,
"aa" :: "ab" :: "ba" :: "bb" :: Nil)
}

test("getPartitionsByFilter: cast(chunk as boolean)=true (not a valid partition predicate)") {
testMetastorePartitionFiltering(
attr("chunk").cast(BooleanType) === true,
20170101 to 20170103,
0 to 23,
"aa" :: "ab" :: "ba" :: "bb" :: Nil)
}

test("getPartitionsByFilter: 20170101=ds") {
testMetastorePartitionFiltering(
Literal(20170101) === attr("ds"),
Expand All @@ -138,7 +154,7 @@ class HiveClientSuite(version: String)
"aa" :: "ab" :: "ba" :: "bb" :: Nil)
}

test("getPartitionsByFilter: chunk in cast(ds as long)=20170101L") {
test("getPartitionsByFilter: cast(ds as long)=20170101L and h=10") {
testMetastorePartitionFiltering(
attr("ds").cast(LongType) === 20170101L && attr("h") === 10,
20170101 to 20170101,
Expand Down