e7af44861e
### What changes were proposed in this pull request?
Apache Parquet 1.12.0 switches its ZSTD compression from Hadoop codec to its own codec.
### Why are the changes needed?
**Apache Spark 3.1 (It requires libhadoop built with zstd)**
```scala
scala> spark.range(10).write.option("compression", "zstd").parquet("/tmp/a")
21/03/27 08:49:38 ERROR Executor: Exception in task 11.0 in stage 0.0 (TID 11)2]
java.lang.RuntimeException: native zStandard library not available:
this version of libhadoop was built without zstd support.
```
**Apache Spark 3.2 (No libhadoop requirement)**
```scala
scala> spark.range(10).write.option("compression", "zstd").parquet("/tmp/a")
```
### Does this PR introduce _any_ user-facing change?
Yes, this is an improvement.
### How was this patch tested?
Pass the CI with the newly added test coverage.
Closes #31981 from dongjoon-hyun/SPARK-34880.
Authored-by: Dongjoon Hyun <dhyun@apple.com>
Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
|
||
|---|---|---|
| .. | ||
| benchmarks | ||
| src | ||
| pom.xml | ||