[
https://issues.apache.org/jira/browse/SPARK-53043?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Dongjoon Hyun updated SPARK-53043:
----------------------------------
Description:
Since Java 9+, we can use Java's built-in `transferTo` directly which is
**significantly faster (over 100x)** than `IOUtils.copy`. In addition, Java's
`transferTo` returns the correct value of copied bytes while `IOUtils.copy`
returns -1 after 2GB which is a well-known limitation.
{code}
scala> import java.io._
import java.io._
scala> spark.time(new FileInputStream("/tmp/4G.bin").transferTo(new
FileOutputStream("/dev/null")))
Time taken: 4 ms
val res0: Long = 4294967296
scala> spark.time(org.apache.commons.io.IOUtils.copy(new
FileInputStream("/tmp/4G.bin"), new FileOutputStream("/dev/null")))
Time taken: 781 ms
val res1: Int = -1
{code}
> Use Java `InputStream.transferTo` instead of `IOUtils.copy`
> -----------------------------------------------------------
>
> Key: SPARK-53043
> URL: https://issues.apache.org/jira/browse/SPARK-53043
> Project: Spark
> Issue Type: Sub-task
> Components: Kubernetes, Spark Core, SQL
> Affects Versions: 4.1.0
> Reporter: Dongjoon Hyun
> Assignee: Dongjoon Hyun
> Priority: Major
> Labels: pull-request-available
> Fix For: 4.1.0
>
>
> Since Java 9+, we can use Java's built-in `transferTo` directly which is
> **significantly faster (over 100x)** than `IOUtils.copy`. In addition, Java's
> `transferTo` returns the correct value of copied bytes while `IOUtils.copy`
> returns -1 after 2GB which is a well-known limitation.
> {code}
> scala> import java.io._
> import java.io._
> scala> spark.time(new FileInputStream("/tmp/4G.bin").transferTo(new
> FileOutputStream("/dev/null")))
> Time taken: 4 ms
> val res0: Long = 4294967296
> scala> spark.time(org.apache.commons.io.IOUtils.copy(new
> FileInputStream("/tmp/4G.bin"), new FileOutputStream("/dev/null")))
> Time taken: 781 ms
> val res1: Int = -1
> {code}
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
