Skip to content

[SPARK-23727][SQL] Support for pushing down filters for DateType in parquet #20851

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
yucai wants to merge 13 commits into from
Closed

[SPARK-23727][SQL] Support for pushing down filters for DateType in parquet #20851

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
13 commits
Select commit Hold shift + click to select a range
079af71
[SPARK-23727][SQL] Support DATE predict push down in parquet
yucai Mar 18, 2018
59d8483
minor, code style
yucai Mar 18, 2018
1f2b450
Add spark.sql.parquet.filterPushdown.date
yucai Mar 18, 2018
4df29df
Make spark.sql.parquet.filterPushdown.date internal
yucai Mar 19, 2018
8f36d1e
Use DateTimeUtils.fromJavaDate
yucai Mar 19, 2018
82c5a73
code style
yucai Mar 20, 2018
784b625
Use 1.date instead of 1.d
yucai Mar 21, 2018
7946bea
Should respect session local timezone
yucai Mar 21, 2018
423a938
Make suite more readable
yucai Mar 22, 2018
a58eec8
code style
yucai Mar 22, 2018
b43e36c
minor
yucai Mar 22, 2018
b70def0
Set spark.sql.parquet.filterPushdown.date=true by default
yucai Mar 26, 2018
759f468
convert date without timezone
yucai Mar 29, 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
9 changes: 9 additions & 0 deletions sql/catalyst/src/main/scala/org/apache/spark/sql/internal/SQLConf.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -353,6 +353,13 @@ object SQLConf {
.booleanConf
.createWithDefault(true)

val PARQUET_FILTER_PUSHDOWN_DATE_ENABLED = buildConf("spark.sql.parquet.filterPushdown.date")
.doc("If true, enables Parquet filter push-down optimization for Date. " +
Copy link
Member

@viirya viirya Mar 19, 2018

Choose a reason for hiding this comment

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

Should be an internal conf, i.e., .internal()?

Copy link
Contributor Author

@yucai yucai Mar 19, 2018

Choose a reason for hiding this comment

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

@gatorsmile any idea?

Copy link
Member

@gatorsmile gatorsmile Mar 23, 2018

Choose a reason for hiding this comment

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

Yes, it should be an internal conf. In Spark 3.0 release, we will revisit all the internal confs and remove all the unnecessary confs.

"This configuration only has an effect when 'spark.sql.parquet.filterPushdown' is enabled.")
.internal()
.booleanConf
.createWithDefault(true)
Copy link
Member

@dongjoon-hyun dongjoon-hyun Mar 22, 2018

Choose a reason for hiding this comment

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

For this kind of thing, we had better start with false.

Copy link
Member

@HyukjinKwon HyukjinKwon Mar 22, 2018

Choose a reason for hiding this comment

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

+1


val PARQUET_WRITE_LEGACY_FORMAT = buildConf("spark.sql.parquet.writeLegacyFormat")
.doc("Whether to be compatible with the legacy Parquet format adopted by Spark 1.4 and prior " +
"versions, when converting Parquet schema to Spark SQL schema and vice versa.")
Expand Down Expand Up @@ -1319,6 +1326,8 @@ class SQLConf extends Serializable with Logging {

def parquetFilterPushDown: Boolean = getConf(PARQUET_FILTER_PUSHDOWN_ENABLED)

def parquetFilterPushDownDate: Boolean = getConf(PARQUET_FILTER_PUSHDOWN_DATE_ENABLED)

def orcFilterPushDown: Boolean = getConf(ORC_FILTER_PUSHDOWN_ENABLED)

def verifyPartitionPath: Boolean = getConf(HIVE_VERIFY_PARTITION_PATH)
Expand Down
33 changes: 33 additions & 0 deletions ...re/src/main/scala/org/apache/spark/sql/execution/datasources/parquet/ParquetFilters.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -17,10 +17,15 @@

package org.apache.spark.sql.execution.datasources.parquet

import java.sql.Date

import org.apache.parquet.filter2.predicate._
import org.apache.parquet.filter2.predicate.FilterApi._
import org.apache.parquet.io.api.Binary

import org.apache.spark.sql.catalyst.util.DateTimeUtils
import org.apache.spark.sql.catalyst.util.DateTimeUtils.SQLDate
import org.apache.spark.sql.internal.SQLConf
import org.apache.spark.sql.sources
import org.apache.spark.sql.types._

Expand All @@ -29,6 +34,10 @@ import org.apache.spark.sql.types._
*/
private[parquet] object ParquetFilters {

private def dateToDays(date: Date): SQLDate = {
DateTimeUtils.fromJavaDate(date)
}

private val makeEq: PartialFunction[DataType, (String, Any) => FilterPredicate] = {
case BooleanType =>
(n: String, v: Any) => FilterApi.eq(booleanColumn(n), v.asInstanceOf[java.lang.Boolean])
Expand All @@ -50,6 +59,10 @@ private[parquet] object ParquetFilters {
(n: String, v: Any) => FilterApi.eq(
binaryColumn(n),
Option(v).map(b => Binary.fromReusedByteArray(v.asInstanceOf[Array[Byte]])).orNull)
case DateType if SQLConf.get.parquetFilterPushDownDate =>
(n: String, v: Any) => FilterApi.eq(
intColumn(n),
Option(v).map(date => dateToDays(date.asInstanceOf[Date]).asInstanceOf[Integer]).orNull)
Copy link
Contributor

@cloud-fan cloud-fan Mar 26, 2018

Choose a reason for hiding this comment

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

sorry I was wrong. I took a look at how these dates get created, in DataSourceStrategy.translateFilter. Actually they are created via DateTimeUtils.toJavaDate without timezone, which means here we should not use timezone either.

}

private val makeNotEq: PartialFunction[DataType, (String, Any) => FilterPredicate] = {
Expand All @@ -72,6 +85,10 @@ private[parquet] object ParquetFilters {
(n: String, v: Any) => FilterApi.notEq(
binaryColumn(n),
Option(v).map(b => Binary.fromReusedByteArray(v.asInstanceOf[Array[Byte]])).orNull)
case DateType if SQLConf.get.parquetFilterPushDownDate =>
(n: String, v: Any) => FilterApi.notEq(
intColumn(n),
Option(v).map(date => dateToDays(date.asInstanceOf[Date]).asInstanceOf[Integer]).orNull)
}

private val makeLt: PartialFunction[DataType, (String, Any) => FilterPredicate] = {
Expand All @@ -91,6 +108,10 @@ private[parquet] object ParquetFilters {
case BinaryType =>
(n: String, v: Any) =>
FilterApi.lt(binaryColumn(n), Binary.fromReusedByteArray(v.asInstanceOf[Array[Byte]]))
case DateType if SQLConf.get.parquetFilterPushDownDate =>
(n: String, v: Any) => FilterApi.lt(
intColumn(n),
Option(v).map(date => dateToDays(date.asInstanceOf[Date]).asInstanceOf[Integer]).orNull)
}

private val makeLtEq: PartialFunction[DataType, (String, Any) => FilterPredicate] = {
Expand All @@ -110,6 +131,10 @@ private[parquet] object ParquetFilters {
case BinaryType =>
(n: String, v: Any) =>
FilterApi.ltEq(binaryColumn(n), Binary.fromReusedByteArray(v.asInstanceOf[Array[Byte]]))
case DateType if SQLConf.get.parquetFilterPushDownDate =>
(n: String, v: Any) => FilterApi.ltEq(
intColumn(n),
Option(v).map(date => dateToDays(date.asInstanceOf[Date]).asInstanceOf[Integer]).orNull)
}

private val makeGt: PartialFunction[DataType, (String, Any) => FilterPredicate] = {
Expand All @@ -129,6 +154,10 @@ private[parquet] object ParquetFilters {
case BinaryType =>
(n: String, v: Any) =>
FilterApi.gt(binaryColumn(n), Binary.fromReusedByteArray(v.asInstanceOf[Array[Byte]]))
case DateType if SQLConf.get.parquetFilterPushDownDate =>
Copy link
Contributor

@cloud-fan cloud-fan Mar 26, 2018

Choose a reason for hiding this comment

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

we should refactor it, so that adding a new data type doesn't need to touch so many places. This can be done later.

(n: String, v: Any) => FilterApi.gt(
intColumn(n),
Option(v).map(date => dateToDays(date.asInstanceOf[Date]).asInstanceOf[Integer]).orNull)
}

private val makeGtEq: PartialFunction[DataType, (String, Any) => FilterPredicate] = {
Expand All @@ -148,6 +177,10 @@ private[parquet] object ParquetFilters {
case BinaryType =>
(n: String, v: Any) =>
FilterApi.gtEq(binaryColumn(n), Binary.fromReusedByteArray(v.asInstanceOf[Array[Byte]]))
case DateType if SQLConf.get.parquetFilterPushDownDate =>
(n: String, v: Any) => FilterApi.gtEq(
intColumn(n),
Option(v).map(date => dateToDays(date.asInstanceOf[Date]).asInstanceOf[Integer]).orNull)
}

/**
Expand Down
50 changes: 47 additions & 3 deletions ...rc/test/scala/org/apache/spark/sql/execution/datasources/parquet/ParquetFilterSuite.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -18,6 +18,7 @@
package org.apache.spark.sql.execution.datasources.parquet

import java.nio.charset.StandardCharsets
import java.sql.Date

import org.apache.parquet.filter2.predicate.{FilterPredicate, Operators}
import org.apache.parquet.filter2.predicate.FilterApi._
Expand Down Expand Up @@ -76,8 +77,10 @@ class ParquetFilterSuite extends QueryTest with ParquetTest with SharedSQLContex
expected: Seq[Row]): Unit = {
val output = predicate.collect { case a: Attribute => a }.distinct

withSQLConf(SQLConf.PARQUET_FILTER_PUSHDOWN_ENABLED.key -> "true") {
withSQLConf(SQLConf.PARQUET_VECTORIZED_READER_ENABLED.key -> "false") {
withSQLConf(
SQLConf.PARQUET_FILTER_PUSHDOWN_ENABLED.key -> "true",
SQLConf.PARQUET_FILTER_PUSHDOWN_DATE_ENABLED.key -> "true",
SQLConf.PARQUET_VECTORIZED_READER_ENABLED.key -> "false") {
val query = df
.select(output.map(e => Column(e)): _*)
.where(Column(predicate))
Expand All @@ -102,7 +105,6 @@ class ParquetFilterSuite extends QueryTest with ParquetTest with SharedSQLContex
maybeFilter.exists(_.getClass === filterClass)
}
checker(stripSparkFilter(query), expected)
}
}
}

Expand Down Expand Up @@ -313,6 +315,48 @@ class ParquetFilterSuite extends QueryTest with ParquetTest with SharedSQLContex
}
}

test("filter pushdown - date") {
implicit class StringToDate(s: String) {
def date: Date = Date.valueOf(s)
}

val data = Seq("2018-03-18", "2018-03-19", "2018-03-20", "2018-03-21")

withParquetDataFrame(data.map(i => Tuple1(i.date))) { implicit df =>
checkFilterPredicate('_1.isNull, classOf[Eq[_]], Seq.empty[Row])
checkFilterPredicate('_1.isNotNull, classOf[NotEq[_]], data.map(i => Row.apply(i.date)))

checkFilterPredicate('_1 === "2018-03-18".date, classOf[Eq[_]], "2018-03-18".date)
checkFilterPredicate('_1 <=> "2018-03-18".date, classOf[Eq[_]], "2018-03-18".date)
checkFilterPredicate('_1 =!= "2018-03-18".date, classOf[NotEq[_]],
Seq("2018-03-19", "2018-03-20", "2018-03-21").map(i => Row.apply(i.date)))

checkFilterPredicate('_1 < "2018-03-19".date, classOf[Lt[_]], "2018-03-18".date)
checkFilterPredicate('_1 > "2018-03-20".date, classOf[Gt[_]], "2018-03-21".date)
checkFilterPredicate('_1 <= "2018-03-18".date, classOf[LtEq[_]], "2018-03-18".date)
checkFilterPredicate('_1 >= "2018-03-21".date, classOf[GtEq[_]], "2018-03-21".date)

checkFilterPredicate(
Literal("2018-03-18".date) === '_1, classOf[Eq[_]], "2018-03-18".date)
checkFilterPredicate(
Literal("2018-03-18".date) <=> '_1, classOf[Eq[_]], "2018-03-18".date)
checkFilterPredicate(
Literal("2018-03-19".date) > '_1, classOf[Lt[_]], "2018-03-18".date)
checkFilterPredicate(
Literal("2018-03-20".date) < '_1, classOf[Gt[_]], "2018-03-21".date)
checkFilterPredicate(
Literal("2018-03-18".date) >= '_1, classOf[LtEq[_]], "2018-03-18".date)
checkFilterPredicate(
Literal("2018-03-21".date) <= '_1, classOf[GtEq[_]], "2018-03-21".date)

checkFilterPredicate(!('_1 < "2018-03-21".date), classOf[GtEq[_]], "2018-03-21".date)
checkFilterPredicate(
'_1 < "2018-03-19".date || '_1 > "2018-03-20".date,
classOf[Operators.Or],
Seq(Row("2018-03-18".date), Row("2018-03-21".date)))
}
}

test("SPARK-6554: don't push down predicates which reference partition columns") {
import testImplicits._

Expand Down