Skip to content

SPARK-1240: handle the case of empty RDD when takeSample #135

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
CodingCat wants to merge 5 commits into from
Closed

SPARK-1240: handle the case of empty RDD when takeSample #135

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
ad483fd
handle the case with empty RDD when take sample
CodingCat Mar 13, 2014
a40e8fb
replace if with require
CodingCat Mar 13, 2014
810948d
further fix
CodingCat Mar 13, 2014
36db06b
create new test cases for takeSample from an empty red
CodingCat Mar 13, 2014
fef57d4
fix the same problem in PySpark
CodingCat Mar 14, 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
9 changes: 8 additions & 1 deletion core/src/main/scala/org/apache/spark/rdd/RDD.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -310,6 +310,9 @@ abstract class RDD[T: ClassTag](
* Return a sampled subset of this RDD.
*/
def sample(withReplacement: Boolean, fraction: Double, seed: Int): RDD[T] = {
if (fraction < Double.MinValue || fraction > Double.MaxValue) {
Copy link
Contributor

@rxin rxin Mar 13, 2014

Choose a reason for hiding this comment

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

Use require. i.e.

require(fraction > Double.MinValue && fraction < Double.MaxValue, "...")

Shouldn't you just check for fraction > 0 but < 1?

Copy link
Contributor

@mengxr mengxr Mar 13, 2014

Choose a reason for hiding this comment

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

The lower bound should be >= 0.0. Sample with replacement can have a faction greater than 1.0.

Copy link
Contributor Author

@CodingCat CodingCat Mar 13, 2014

Choose a reason for hiding this comment

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

Hi, @rxin , I'm also a bit confused here, I think the name of the argument is a bit confusing

https://github.com/apache/spark/blob/master/core/src/main/scala/org/apache/spark/rdd/RDD.scala#L357

The above line contains a multiplier to ensure that the sampling can return enough sample points in most of cases..(I think so), so the fraction value can actually be larger than 1

also, this value actually determines the mean value of Poisson/Bernoulli distribution

https://github.com/apache/spark/blob/master/core/src/main/scala/org/apache/spark/rdd/RDD.scala#L314

throw new Exception("Invalid fraction value:" + fraction)
}
if (withReplacement) {
new PartitionwiseSampledRDD[T, T](this, new PoissonSampler[T](fraction), seed)
} else {
Expand Down Expand Up @@ -344,6 +347,10 @@ abstract class RDD[T: ClassTag](
throw new IllegalArgumentException("Negative number of elements requested")
}

if (initialCount == 0) {
return new Array[T](0)
}

if (initialCount > Integer.MAX_VALUE - 1) {
maxSelected = Integer.MAX_VALUE - 1
} else {
Expand All @@ -362,7 +369,7 @@ abstract class RDD[T: ClassTag](
var samples = this.sample(withReplacement, fraction, rand.nextInt()).collect()

// If the first sample didn't turn out large enough, keep trying to take samples;
// this shouldn't happen often because we use a big multiplier for thei initial size
// this shouldn't happen often because we use a big multiplier for the initial size
while (samples.length < total) {
samples = this.sample(withReplacement, fraction, rand.nextInt()).collect()
}
Expand Down
4 changes: 4 additions & 0 deletions core/src/test/scala/org/apache/spark/rdd/RDDSuite.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -457,6 +457,10 @@ class RDDSuite extends FunSuite with SharedSparkContext {

test("takeSample") {
val data = sc.parallelize(1 to 100, 2)
val emptySet = data.filter(_ => false)
Copy link
Contributor

@mengxr mengxr Mar 13, 2014

Choose a reason for hiding this comment

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

Is there a better way to create an empty RDD?

Copy link
Contributor

@rxin rxin Mar 13, 2014

Choose a reason for hiding this comment

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

yup do

data.mapPartitions { iter => Iterator.empty }


val sample = emptySet.takeSample(false, 20, 1)
assert(sample.size === 0)
for (seed <- 1 to 5) {
val sample = data.takeSample(withReplacement=false, 20, seed)
assert(sample.size === 20) // Got exactly 20 elements
Expand Down