ODIn/spark-instrumented-optimizer
1
0
Fork 0
Code Issues Pull requests Projects Releases Wiki Activity

[SPARK-34954][SQL] Use zstd codec name in ORC file names

Browse source 
### What changes were proposed in this pull request?

This PR aims to add `zstd` codec names in the Spark generated ORC file names for consistency.

### Why are the changes needed?

Like the other ORC supported codecs, we had better have `zstd` in the Spark generated ORC file names. Please note that there is no problem at reading/writing ORC zstd files currently. This PR only aims to revise the file name format for consistency.

**SNAPPY**
```
scala> spark.range(10).repartition(1).write.option("compression", "snappy").orc("/tmp/snappy")

$ ls -al /tmp/snappy
total 24
drwxr-xr-x   6 dongjoon  wheel  192 Apr  4 12:17 .
drwxrwxrwt  14 root      wheel  448 Apr  4 12:17 ..
-rw-r--r--   1 dongjoon  wheel    8 Apr  4 12:17 ._SUCCESS.crc
-rw-r--r--   1 dongjoon  wheel   12 Apr  4 12:17 .part-00000-833bb7ad-d1e1-48cc-9719-07b2d594aa4c-c000.snappy.orc.crc
-rw-r--r--   1 dongjoon  wheel    0 Apr  4 12:17 _SUCCESS
-rw-r--r--   1 dongjoon  wheel  231 Apr  4 12:17 part-00000-833bb7ad-d1e1-48cc-9719-07b2d594aa4c-c000.snappy.orc
```

**ZSTD (AS-IS)**
```
scala> spark.range(10).repartition(1).write.option("compression", "zstd").orc("/tmp/zstd")

$ ls -al /tmp/zstd
total 24
drwxr-xr-x   6 dongjoon  wheel  192 Apr  4 12:17 .
drwxrwxrwt  14 root      wheel  448 Apr  4 12:17 ..
-rw-r--r--   1 dongjoon  wheel    8 Apr  4 12:17 ._SUCCESS.crc
-rw-r--r--   1 dongjoon  wheel   12 Apr  4 12:17 .part-00000-2f403ce9-7314-4db5-bca3-b1c1dd83335f-c000.orc.crc
-rw-r--r--   1 dongjoon  wheel    0 Apr  4 12:17 _SUCCESS
-rw-r--r--   1 dongjoon  wheel  231 Apr  4 12:17 part-00000-2f403ce9-7314-4db5-bca3-b1c1dd83335f-c000.orc
```

**ZSTD (After this PR)**
```
scala> spark.range(10).repartition(1).write.option("compression", "zstd").orc("/tmp/zstd_new")

$ ls -al /tmp/zstd_new
total 24
drwxr-xr-x   6 dongjoon  wheel  192 Apr  4 12:28 .
drwxrwxrwt  15 root      wheel  480 Apr  4 12:28 ..
-rw-r--r--   1 dongjoon  wheel    8 Apr  4 12:28 ._SUCCESS.crc
-rw-r--r--   1 dongjoon  wheel   12 Apr  4 12:28 .part-00000-49d57329-7196-4caf-839c-4251c876e26b-c000.zstd.orc.crc
-rw-r--r--   1 dongjoon  wheel    0 Apr  4 12:28 _SUCCESS
-rw-r--r--   1 dongjoon  wheel  231 Apr  4 12:28 part-00000-49d57329-7196-4caf-839c-4251c876e26b-c000.zstd.orc
```

### Does this PR introduce _any_ user-facing change?

No.

### How was this patch tested?

Pass the CIs with the updated UT.

Closes #32051 from dongjoon-hyun/SPARK-34954.

Authored-by: Dongjoon Hyun <dhyun@apple.com>
Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
This commit is contained in:
Dongjoon Hyun 2021-04-04 17:11:56 -07:00
parent ba32b200e4
commit 748f05fca9
2 changed files with 3 additions and 0 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

1
sql/core/src/main/scala/org/apache/spark/sql/execution/datasources/orc/OrcUtils.scala
View file

@ -45,6 +45,7 @@ object OrcUtils extends Logging {
"NONE" -> "",
"SNAPPY" -> ".snappy",
"ZLIB" -> ".zlib",
"ZSTD" -> ".zstd",
"LZO" -> ".lzo")
def listOrcFiles(pathStr: String, conf: Configuration): Seq[Path] = {

2
sql/core/src/test/scala/org/apache/spark/sql/execution/datasources/orc/OrcSourceSuite.scala
View file

@ -602,6 +602,8 @@ class OrcSourceSuite extends OrcSuite with SharedSparkSession {
val path = dir.getAbsolutePath
spark.range(3).write.option("compression", "zstd").orc(path)
checkAnswer(spark.read.orc(path), Seq(Row(0), Row(1), Row(2)))
val files = OrcUtils.listOrcFiles(path, spark.sessionState.newHadoopConf())
assert(files.nonEmpty && files.forall(_.getName.contains("zstd")))
}
}