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

[SPARK-4453][SPARK-4213][SQL] Simplifies Parquet filter generation code #3317

Closed
wants to merge 2 commits into from
+161 −693
Closed

[SPARK-4453][SPARK-4213][SQL] Simplifies Parquet filter generation code #3317

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
43760e8
Simplifies Parquet filter generation logic
liancheng Nov 17, 2014
d6a9499
Fixes import styling issue
liancheng Nov 17, 2014
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
1 change: 1 addition & 0 deletions sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/expressions/namedExpressions.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -26,6 +26,7 @@ import org.apache.spark.sql.catalyst.util.Metadata
object NamedExpression {
private val curId = new java.util.concurrent.atomic.AtomicLong()
def newExprId = ExprId(curId.getAndIncrement())
def unapply(expr: NamedExpression): Option[(String, DataType)] = Some(expr.name, expr.dataType)
Copy link
Contributor

Choose a reason for hiding this comment

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

Cute. At some point we should probably write up a guide of all the ways we use pattern matching in catalyst, perhaps as a part of your big refactoring PR. I'm generally in support of these shortcuts, but I want to make sure that we are coherent about how they are used. BTW, I've been holding off on merging your other PR because I didn't want to create conflicts close to the release, but I would like to revisit this soon .

}

/**
Expand Down
25 changes: 9 additions & 16 deletions sql/core/src/main/scala/org/apache/spark/sql/execution/SparkStrategies.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -209,22 +209,15 @@ private[sql] abstract class SparkStrategies extends QueryPlanner[SparkPlan] {
case PhysicalOperation(projectList, filters: Seq[Expression], relation: ParquetRelation) =>
val prunePushedDownFilters =
if (sqlContext.parquetFilterPushDown) {
(filters: Seq[Expression]) => {
filters.filter { filter =>
// Note: filters cannot be pushed down to Parquet if they contain more complex
// expressions than simple "Attribute cmp Literal" comparisons. Here we remove
// all filters that have been pushed down. Note that a predicate such as
// "(A AND B) OR C" can result in "A OR C" being pushed down.
val recordFilter = ParquetFilters.createFilter(filter)
if (!recordFilter.isDefined) {
// First case: the pushdown did not result in any record filter.
true
} else {
// Second case: a record filter was created; here we are conservative in
// the sense that even if "A" was pushed and we check for "A AND B" we
// still want to keep "A AND B" in the higher-level filter, not just "B".
!ParquetFilters.findExpression(recordFilter.get, filter).isDefined
}
(predicates: Seq[Expression]) => {
// Note: filters cannot be pushed down to Parquet if they contain more complex
// expressions than simple "Attribute cmp Literal" comparisons. Here we remove all
// filters that have been pushed down. Note that a predicate such as "(A AND B) OR C"
// can result in "A OR C" being pushed down. Here we are conservative in the sense
// that even if "A" was pushed and we check for "A AND B" we still want to keep
// "A AND B" in the higher-level filter, not just "B".
predicates.map(p => p -> ParquetFilters.createFilter(p)).collect {
case (predicate, None) => predicate
}
}
} else {
Expand Down
Loading