deniskuzZ commented on code in PR #5076:
URL: https://github.com/apache/hive/pull/5076#discussion_r1500832533
##########
iceberg/iceberg-handler/src/main/java/org/apache/iceberg/mr/hive/HiveIcebergStorageHandler.java:
##########
@@ -2021,4 +2023,34 @@ public List<Partition>
getPartitionsByExpr(org.apache.hadoop.hive.ql.metadata.Ta
throw new SemanticException(String.format("Error while fetching the
partitions due to: %s", e));
}
}
+
+ @Override
+ public boolean supportsMergeFiles() {
+ return true;
+ }
+
+ @Override
+ public List<FileStatus> getMergeInputFiles(Properties properties) throws
IOException {
+ String tableName = properties.getProperty(Catalogs.NAME);
+ String snapshotRef = properties.getProperty(Catalogs.SNAPSHOT_REF);
+ Configuration configuration = SessionState.getSessionConf();
+ List<JobContext> originalContextList = generateJobContext(configuration,
tableName, snapshotRef);
+ List<JobContext> jobContextList = originalContextList.stream()
+ .map(TezUtil::enrichContextWithVertexId)
+ .collect(Collectors.toList());
+ if (originalContextList.isEmpty()) {
+ return Collections.emptyList();
+ }
+ List<FileStatus> dataFiles = Lists.newArrayList();
+ for (DataFile dataFile : new
HiveIcebergOutputCommitter().getWrittenFiles(jobContextList)) {
Review Comment:
1. why do you need new instance of HiveIcebergOutputCommitter, can't you
make the method static? Maybe we should move it to Iceberg util class
2. Why do you need to enrichContextWithVertexId again inside of
`getWrittenFiles`?
3. Why don't we return `List<FileStatus> dataFiles` right away, but add 1
more foreach with transformation?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]