[
https://issues.apache.org/jira/browse/PARQUET-2380?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17787234#comment-17787234
]
ASF GitHub Bot commented on PARQUET-2380:
-----------------------------------------
amousavigourabi commented on code in PR #1195:
URL: https://github.com/apache/parquet-mr/pull/1195#discussion_r1397394681
##########
parquet-hadoop/src/main/java/org/apache/parquet/hadoop/rewrite/RewriteOptions.java:
##########
@@ -64,15 +72,53 @@ private RewriteOptions(Configuration conf,
this.indexCacheStrategy = indexCacheStrategy;
}
+ /**
+ * Gets the {@link Configuration} part of the rewrite options.
+ *
+ * @return the associated {@link Configuration}
+ */
public Configuration getConf() {
+ return ConfigurationUtil.createHadoopConfiguration(conf);
+ }
+
+ /**
+ * Gets the {@link ParquetConfiguration} part of the rewrite options.
+ *
+ * @return the associated {@link ParquetConfiguration}
+ */
+ public ParquetConfiguration getParquetConfiguration() {
return conf;
}
- public List<Path> getInputFiles() {
Review Comment:
It seems japicmp didn't pick it up as the erasure is the same as in
`List<InputFile> getInputFiles()`. I'll restore the method.
> Decouple RewriteOptions from Hadoop classes
> -------------------------------------------
>
> Key: PARQUET-2380
> URL: https://issues.apache.org/jira/browse/PARQUET-2380
> Project: Parquet
> Issue Type: Improvement
> Components: parquet-mr
> Reporter: Atour Mousavi Gourabi
> Priority: Major
>
> ParquetRewriter's RewriteOptions makes use of Hadoop Path and Configuration,
> where it could instead allow users to specify these using the Parquet
> interface methods as well. This would allow for proper decoupling Hadoop for
> rewriting in a later stage as well.
> This is part of a larger effort to decouple Parquet from Hadoop libraries.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
