Skip to content

[SPARK-19761][SQL]create InMemoryFileIndex with an empty rootPaths when set PARALLEL_PARTITION_DISCOVERY_THRESHOLD to zero failed #17093

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
windpiger wants to merge 9 commits into from
Closed

[SPARK-19761][SQL]create InMemoryFileIndex with an empty rootPaths when set PARALLEL_PARTITION_DISCOVERY_THRESHOLD to zero failed #17093

Show file tree
Hide file tree
Changes from 8 commits
Commits
Show all changes
9 commits
Select commit Hold shift + click to select a range
96898a2
[SPARK-19761][SQL]create InMemoryFileIndex with an empty rootPaths wh…
windpiger Feb 28, 2017
a3ac29b
cover the negtive number situation
windpiger Feb 28, 2017
ec0afac
merge with master
windpiger Feb 28, 2017
13b70f0
add checkValue for the conf
windpiger Mar 1, 2017
1cb997c
fix comments
windpiger Mar 1, 2017
0d2334d
fix comments
windpiger Mar 1, 2017
74d08a5
modify test case
windpiger Mar 1, 2017
e1a9072
modify a test name
windpiger Mar 1, 2017
3c079ff
modify a style
windpiger Mar 1, 2017
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
2 changes: 1 addition & 1 deletion ...rc/main/scala/org/apache/spark/sql/execution/datasources/PartitioningAwareFileIndex.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -300,7 +300,7 @@ object PartitioningAwareFileIndex extends Logging {
sparkSession: SparkSession): Seq[(Path, Seq[FileStatus])] = {

// Short-circuits parallel listing when serial listing is likely to be faster.
if (paths.size < sparkSession.sessionState.conf.parallelPartitionDiscoveryThreshold) {
if (paths.size <= sparkSession.sessionState.conf.parallelPartitionDiscoveryThreshold) {
Copy link
Contributor

@cloud-fan cloud-fan Feb 28, 2017

Choose a reason for hiding this comment

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

why not we just make sure parallelPartitionDiscoveryThreshold is greater than 0? We can add a condition(via checkValue) in SQLConf.PARALLEL_PARTITION_DISCOVERY_THRESHOLD

Copy link
Contributor Author

@windpiger windpiger Mar 1, 2017

Choose a reason for hiding this comment

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

sorry, I didn't notice there is a checkValue func, let me fix it. thanks!

Copy link
Contributor Author

@windpiger windpiger Mar 1, 2017

Choose a reason for hiding this comment

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

add equal is more clear to understand the conf parallelPartitionDiscoveryThreshold

return paths.map { path =>
(path, listLeafFiles(path, hadoopConf, filter, Some(sparkSession)))
}
Expand Down
6 changes: 4 additions & 2 deletions sql/core/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 @@ -402,11 +402,13 @@ object SQLConf {

val PARALLEL_PARTITION_DISCOVERY_THRESHOLD =
buildConf("spark.sql.sources.parallelPartitionDiscovery.threshold")
.doc("The maximum number of files allowed for listing files at driver side. If the number " +
"of detected files exceeds this value during partition discovery, it tries to list the " +
.doc("The maximum number of paths allowed for listing files at driver side. If the number " +
"of detected paths exceeds this value during partition discovery, it tries to list the " +
"files with another Spark distributed job. This applies to Parquet, ORC, CSV, JSON and " +
"LibSVM data sources.")
.intConf
.checkValue(parallel => parallel >= 0, "The maximum number of paths allowed for listing " +
"files at driver side must not be negative")
.createWithDefault(32)

val PARALLEL_PARTITION_DISCOVERY_PARALLELISM =
Expand Down
16 changes: 16 additions & 0 deletions sql/core/src/test/scala/org/apache/spark/sql/execution/datasources/FileIndexSuite.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -27,6 +27,7 @@ import org.apache.hadoop.fs.{FileStatus, Path, RawLocalFileSystem}

import org.apache.spark.metrics.source.HiveCatalogMetrics
import org.apache.spark.sql.catalyst.util._
import org.apache.spark.sql.internal.SQLConf
import org.apache.spark.sql.test.SharedSQLContext

class FileIndexSuite extends SharedSQLContext {
Expand Down Expand Up @@ -179,6 +180,21 @@ class FileIndexSuite extends SharedSQLContext {
}
}

test("InMemoryFileIndex with empty rootPaths when PARALLEL_PARTITION_DISCOVERY_THRESHOLD" +
"is a nonpositive number") {
withSQLConf(SQLConf.PARALLEL_PARTITION_DISCOVERY_THRESHOLD.key -> "0") {
new InMemoryFileIndex(spark, Seq.empty, Map.empty, None)
}

val e = intercept[IllegalArgumentException] {
withSQLConf(SQLConf.PARALLEL_PARTITION_DISCOVERY_THRESHOLD.key -> "-1") {
new InMemoryFileIndex(spark, Seq.empty, Map.empty, None)
}
}.getMessage
assert(e.contains("The maximum number of paths allowed for listing files at " +
"driver side must not be negative"))
Copy link
Member

@gatorsmile gatorsmile Mar 1, 2017

Choose a reason for hiding this comment

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

Nit: indent. : )

Copy link
Contributor Author

@windpiger windpiger Mar 1, 2017 •
edited
Loading

Choose a reason for hiding this comment

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

oh...thanks~

}

test("refresh for InMemoryFileIndex with FileStatusCache") {
withTempDir { dir =>
val fileStatusCache = FileStatusCache.getOrCreate(spark)
Expand Down