spark git commit: [SPARK-18853][SQL] Project (UnaryNode) is way too aggressive in estimating statistics

2016-12-14 Thread hvanhovell 
Repository: spark
Updated Branches:
  refs/heads/branch-2.0 1d5c7f452 -> 2b18bd4b9


[SPARK-18853][SQL] Project (UnaryNode) is way too aggressive in estimating 
statistics

This patch reduces the default number element estimation for arrays and maps 
from 100 to 1. The issue with the 100 number is that when nested (e.g. an array 
of map), 100 * 100 would be used as the default size. This sounds like just an 
overestimation which doesn't seem that bad (since it is usually better to 
overestimate than underestimate). However, due to the way we assume the size 
output for Project (new estimated column size / old estimated column size), 
this overestimation can become underestimation. It is actually in general in 
this case safer to assume 1 default element.

This should be covered by existing tests.

Author: Reynold Xin 

Closes #16274 from rxin/SPARK-18853.

(cherry picked from commit 5d799473696a15fddd54ec71a93b6f8cb169810c)
Signed-off-by: Herman van Hovell 
(cherry picked from commit e8866f9fc62095b78421d461549f7eaf8e9070b3)
Signed-off-by: Herman van Hovell 


Project: http://git-wip-us.apache.org/repos/asf/spark/repo
Commit: http://git-wip-us.apache.org/repos/asf/spark/commit/2b18bd4b
Tree: http://git-wip-us.apache.org/repos/asf/spark/tree/2b18bd4b
Diff: http://git-wip-us.apache.org/repos/asf/spark/diff/2b18bd4b

Branch: refs/heads/branch-2.0
Commit: 2b18bd4b99c0c85f28bd174751f52f940f78d2e8
Parents: 1d5c7f4
Author: Reynold Xin 
Authored: Wed Dec 14 21:22:49 2016 +0100
Committer: Herman van Hovell 
Committed: Wed Dec 14 21:25:07 2016 +0100

--
 .../scala/org/apache/spark/sql/types/ArrayType.scala  |  6 +++---
 .../scala/org/apache/spark/sql/types/MapType.scala|  6 +++---
 .../org/apache/spark/sql/types/DataTypeSuite.scala| 14 +++---
 3 files changed, 13 insertions(+), 13 deletions(-)
--


http://git-wip-us.apache.org/repos/asf/spark/blob/2b18bd4b/sql/catalyst/src/main/scala/org/apache/spark/sql/types/ArrayType.scala
--
diff --git 
a/sql/catalyst/src/main/scala/org/apache/spark/sql/types/ArrayType.scala 
b/sql/catalyst/src/main/scala/org/apache/spark/sql/types/ArrayType.scala
index 82a03b0..8c4df58 100644
--- a/sql/catalyst/src/main/scala/org/apache/spark/sql/types/ArrayType.scala
+++ b/sql/catalyst/src/main/scala/org/apache/spark/sql/types/ArrayType.scala
@@ -70,10 +70,10 @@ case class ArrayType(elementType: DataType, containsNull: 
Boolean) extends DataT
   ("containsNull" -> containsNull)
 
   /**
-   * The default size of a value of the ArrayType is 100 * the default size of 
the element type.
-   * (We assume that there are 100 elements).
+   * The default size of a value of the ArrayType is the default size of the 
element type.
+   * We assume that there is only 1 element on average in an array. See 
SPARK-18853.
*/
-  override def defaultSize: Int = 100 * elementType.defaultSize
+  override def defaultSize: Int = 1 * elementType.defaultSize
 
   override def simpleString: String = s"array<${elementType.simpleString}>"
 

http://git-wip-us.apache.org/repos/asf/spark/blob/2b18bd4b/sql/catalyst/src/main/scala/org/apache/spark/sql/types/MapType.scala
--
diff --git 
a/sql/catalyst/src/main/scala/org/apache/spark/sql/types/MapType.scala 
b/sql/catalyst/src/main/scala/org/apache/spark/sql/types/MapType.scala
index 1789609..4d59920 100644
--- a/sql/catalyst/src/main/scala/org/apache/spark/sql/types/MapType.scala
+++ b/sql/catalyst/src/main/scala/org/apache/spark/sql/types/MapType.scala
@@ -57,10 +57,10 @@ case class MapType(
 
   /**
* The default size of a value of the MapType is
-   * 100 * (the default size of the key type + the default size of the value 
type).
-   * (We assume that there are 100 elements).
+   * (the default size of the key type + the default size of the value type).
+   * We assume that there is only 1 element on average in a map. See 
SPARK-18853.
*/
-  override def defaultSize: Int = 100 * (keyType.defaultSize + 
valueType.defaultSize)
+  override def defaultSize: Int = 1 * (keyType.defaultSize + 
valueType.defaultSize)
 
   override def simpleString: String = 
s"map<${keyType.simpleString},${valueType.simpleString}>"
 

http://git-wip-us.apache.org/repos/asf/spark/blob/2b18bd4b/sql/catalyst/src/test/scala/org/apache/spark/sql/types/DataTypeSuite.scala
--
diff --git 
a/sql/catalyst/src/test/scala/org/apache/spark/sql/types/DataTypeSuite.scala 
b/sql/catalyst/src/test/scala/org/apache/spark/sql/types/DataTypeSuite.scala
index 569230a..8c947ed 100644
---

spark git commit: [SPARK-18853][SQL] Project (UnaryNode) is way too aggressive in estimating statistics

2016-12-14 Thread hvanhovell 
Repository: spark
Updated Branches:
  refs/heads/branch-2.1 af12a21ca -> e8866f9fc


[SPARK-18853][SQL] Project (UnaryNode) is way too aggressive in estimating 
statistics

## What changes were proposed in this pull request?
This patch reduces the default number element estimation for arrays and maps 
from 100 to 1. The issue with the 100 number is that when nested (e.g. an array 
of map), 100 * 100 would be used as the default size. This sounds like just an 
overestimation which doesn't seem that bad (since it is usually better to 
overestimate than underestimate). However, due to the way we assume the size 
output for Project (new estimated column size / old estimated column size), 
this overestimation can become underestimation. It is actually in general in 
this case safer to assume 1 default element.

## How was this patch tested?
This should be covered by existing tests.

Author: Reynold Xin 

Closes #16274 from rxin/SPARK-18853.

(cherry picked from commit 5d799473696a15fddd54ec71a93b6f8cb169810c)
Signed-off-by: Herman van Hovell 


Project: http://git-wip-us.apache.org/repos/asf/spark/repo
Commit: http://git-wip-us.apache.org/repos/asf/spark/commit/e8866f9f
Tree: http://git-wip-us.apache.org/repos/asf/spark/tree/e8866f9f
Diff: http://git-wip-us.apache.org/repos/asf/spark/diff/e8866f9f

Branch: refs/heads/branch-2.1
Commit: e8866f9fc62095b78421d461549f7eaf8e9070b3
Parents: af12a21
Author: Reynold Xin 
Authored: Wed Dec 14 21:22:49 2016 +0100
Committer: Herman van Hovell 
Committed: Wed Dec 14 21:23:01 2016 +0100

--
 .../scala/org/apache/spark/sql/types/ArrayType.scala  |  6 +++---
 .../scala/org/apache/spark/sql/types/MapType.scala|  6 +++---
 .../org/apache/spark/sql/types/DataTypeSuite.scala| 14 +++---
 3 files changed, 13 insertions(+), 13 deletions(-)
--


http://git-wip-us.apache.org/repos/asf/spark/blob/e8866f9f/sql/catalyst/src/main/scala/org/apache/spark/sql/types/ArrayType.scala
--
diff --git 
a/sql/catalyst/src/main/scala/org/apache/spark/sql/types/ArrayType.scala 
b/sql/catalyst/src/main/scala/org/apache/spark/sql/types/ArrayType.scala
index d409271..98efba1 100644
--- a/sql/catalyst/src/main/scala/org/apache/spark/sql/types/ArrayType.scala
+++ b/sql/catalyst/src/main/scala/org/apache/spark/sql/types/ArrayType.scala
@@ -78,10 +78,10 @@ case class ArrayType(elementType: DataType, containsNull: 
Boolean) extends DataT
   ("containsNull" -> containsNull)
 
   /**
-   * The default size of a value of the ArrayType is 100 * the default size of 
the element type.
-   * (We assume that there are 100 elements).
+   * The default size of a value of the ArrayType is the default size of the 
element type.
+   * We assume that there is only 1 element on average in an array. See 
SPARK-18853.
*/
-  override def defaultSize: Int = 100 * elementType.defaultSize
+  override def defaultSize: Int = 1 * elementType.defaultSize
 
   override def simpleString: String = s"array<${elementType.simpleString}>"
 

http://git-wip-us.apache.org/repos/asf/spark/blob/e8866f9f/sql/catalyst/src/main/scala/org/apache/spark/sql/types/MapType.scala
--
diff --git 
a/sql/catalyst/src/main/scala/org/apache/spark/sql/types/MapType.scala 
b/sql/catalyst/src/main/scala/org/apache/spark/sql/types/MapType.scala
index fbf3a61..6691b81 100644
--- a/sql/catalyst/src/main/scala/org/apache/spark/sql/types/MapType.scala
+++ b/sql/catalyst/src/main/scala/org/apache/spark/sql/types/MapType.scala
@@ -56,10 +56,10 @@ case class MapType(
 
   /**
* The default size of a value of the MapType is
-   * 100 * (the default size of the key type + the default size of the value 
type).
-   * (We assume that there are 100 elements).
+   * (the default size of the key type + the default size of the value type).
+   * We assume that there is only 1 element on average in a map. See 
SPARK-18853.
*/
-  override def defaultSize: Int = 100 * (keyType.defaultSize + 
valueType.defaultSize)
+  override def defaultSize: Int = 1 * (keyType.defaultSize + 
valueType.defaultSize)
 
   override def simpleString: String = 
s"map<${keyType.simpleString},${valueType.simpleString}>"
 

http://git-wip-us.apache.org/repos/asf/spark/blob/e8866f9f/sql/catalyst/src/test/scala/org/apache/spark/sql/types/DataTypeSuite.scala
--
diff --git 
a/sql/catalyst/src/test/scala/org/apache/spark/sql/types/DataTypeSuite.scala 
b/sql/catalyst/src/test/scala/org/apache/spark/sql/types/DataTypeSuite.scala
index b8ab9a9..12d2c00 100644
--- a/sql/catalyst/src/test/scala/org/apache/spark/sql/types/DataTypeSuite.scala
+++

spark git commit: [SPARK-18853][SQL] Project (UnaryNode) is way too aggressive in estimating statistics

2016-12-14 Thread hvanhovell 
Repository: spark
Updated Branches:
  refs/heads/master 89ae26dcd -> 5d7994736


[SPARK-18853][SQL] Project (UnaryNode) is way too aggressive in estimating 
statistics

## What changes were proposed in this pull request?
This patch reduces the default number element estimation for arrays and maps 
from 100 to 1. The issue with the 100 number is that when nested (e.g. an array 
of map), 100 * 100 would be used as the default size. This sounds like just an 
overestimation which doesn't seem that bad (since it is usually better to 
overestimate than underestimate). However, due to the way we assume the size 
output for Project (new estimated column size / old estimated column size), 
this overestimation can become underestimation. It is actually in general in 
this case safer to assume 1 default element.

## How was this patch tested?
This should be covered by existing tests.

Author: Reynold Xin 

Closes #16274 from rxin/SPARK-18853.


Project: http://git-wip-us.apache.org/repos/asf/spark/repo
Commit: http://git-wip-us.apache.org/repos/asf/spark/commit/5d799473
Tree: http://git-wip-us.apache.org/repos/asf/spark/tree/5d799473
Diff: http://git-wip-us.apache.org/repos/asf/spark/diff/5d799473

Branch: refs/heads/master
Commit: 5d799473696a15fddd54ec71a93b6f8cb169810c
Parents: 89ae26d
Author: Reynold Xin 
Authored: Wed Dec 14 21:22:49 2016 +0100
Committer: Herman van Hovell 
Committed: Wed Dec 14 21:22:49 2016 +0100

--
 .../scala/org/apache/spark/sql/types/ArrayType.scala  |  6 +++---
 .../scala/org/apache/spark/sql/types/MapType.scala|  6 +++---
 .../org/apache/spark/sql/types/DataTypeSuite.scala| 14 +++---
 3 files changed, 13 insertions(+), 13 deletions(-)
--


http://git-wip-us.apache.org/repos/asf/spark/blob/5d799473/sql/catalyst/src/main/scala/org/apache/spark/sql/types/ArrayType.scala
--
diff --git 
a/sql/catalyst/src/main/scala/org/apache/spark/sql/types/ArrayType.scala 
b/sql/catalyst/src/main/scala/org/apache/spark/sql/types/ArrayType.scala
index d409271..98efba1 100644
--- a/sql/catalyst/src/main/scala/org/apache/spark/sql/types/ArrayType.scala
+++ b/sql/catalyst/src/main/scala/org/apache/spark/sql/types/ArrayType.scala
@@ -78,10 +78,10 @@ case class ArrayType(elementType: DataType, containsNull: 
Boolean) extends DataT
   ("containsNull" -> containsNull)
 
   /**
-   * The default size of a value of the ArrayType is 100 * the default size of 
the element type.
-   * (We assume that there are 100 elements).
+   * The default size of a value of the ArrayType is the default size of the 
element type.
+   * We assume that there is only 1 element on average in an array. See 
SPARK-18853.
*/
-  override def defaultSize: Int = 100 * elementType.defaultSize
+  override def defaultSize: Int = 1 * elementType.defaultSize
 
   override def simpleString: String = s"array<${elementType.simpleString}>"
 

http://git-wip-us.apache.org/repos/asf/spark/blob/5d799473/sql/catalyst/src/main/scala/org/apache/spark/sql/types/MapType.scala
--
diff --git 
a/sql/catalyst/src/main/scala/org/apache/spark/sql/types/MapType.scala 
b/sql/catalyst/src/main/scala/org/apache/spark/sql/types/MapType.scala
index fbf3a61..6691b81 100644
--- a/sql/catalyst/src/main/scala/org/apache/spark/sql/types/MapType.scala
+++ b/sql/catalyst/src/main/scala/org/apache/spark/sql/types/MapType.scala
@@ -56,10 +56,10 @@ case class MapType(
 
   /**
* The default size of a value of the MapType is
-   * 100 * (the default size of the key type + the default size of the value 
type).
-   * (We assume that there are 100 elements).
+   * (the default size of the key type + the default size of the value type).
+   * We assume that there is only 1 element on average in a map. See 
SPARK-18853.
*/
-  override def defaultSize: Int = 100 * (keyType.defaultSize + 
valueType.defaultSize)
+  override def defaultSize: Int = 1 * (keyType.defaultSize + 
valueType.defaultSize)
 
   override def simpleString: String = 
s"map<${keyType.simpleString},${valueType.simpleString}>"
 

http://git-wip-us.apache.org/repos/asf/spark/blob/5d799473/sql/catalyst/src/test/scala/org/apache/spark/sql/types/DataTypeSuite.scala
--
diff --git 
a/sql/catalyst/src/test/scala/org/apache/spark/sql/types/DataTypeSuite.scala 
b/sql/catalyst/src/test/scala/org/apache/spark/sql/types/DataTypeSuite.scala
index b8ab9a9..12d2c00 100644
--- a/sql/catalyst/src/test/scala/org/apache/spark/sql/types/DataTypeSuite.scala
+++ b/sql/catalyst/src/test/scala/org/apache/spark/sql/types/DataTypeSuite.scala
@@ -253,7 +253,7 @@ class DataTypeSuite extends SparkFunSuite {