szehon-ho commented on code in PR #7565:
URL: https://github.com/apache/iceberg/pull/7565#discussion_r1190392156
##########
spark/v3.4/spark/src/main/java/org/apache/iceberg/spark/actions/RewritePositionDeleteFilesSparkAction.java:
##########
@@ -133,55 +133,69 @@ public RewritePositionDeleteFiles.Result execute() {
}
}
- private Map<StructLike, List<List<PositionDeletesScanTask>>>
planFileGroups() {
+ private StructLikeMap<List<List<PositionDeletesScanTask>>> planFileGroups() {
Table deletesTable =
MetadataTableUtils.createMetadataTableInstance(table,
MetadataTableType.POSITION_DELETES);
- CloseableIterable<PositionDeletesScanTask> scanTasks =
+
+ CloseableIterable<PositionDeletesScanTask> tasks =
CloseableIterable.transform(
deletesTable.newBatchScan().ignoreResiduals().planFiles(),
t -> (PositionDeletesScanTask) t);
try {
StructType partitionType = Partitioning.partitionType(table);
- StructLikeMap<List<PositionDeletesScanTask>> filesByPartition =
- StructLikeMap.create(partitionType);
-
- for (PositionDeletesScanTask task : scanTasks) {
- StructLike coerced = coercePartition(task, partitionType);
-
- List<PositionDeletesScanTask> partitionTasks =
filesByPartition.get(coerced);
- if (partitionTasks == null) {
- partitionTasks = Lists.newArrayList();
- }
- partitionTasks.add(task);
- filesByPartition.put(coerced, partitionTasks);
- }
-
- StructLikeMap<List<List<PositionDeletesScanTask>>> fileGroupsByPartition
=
- StructLikeMap.create(partitionType);
- filesByPartition.forEach(
- (partition, partitionTasks) -> {
- Iterable<List<PositionDeletesScanTask>> plannedFileGroups =
- rewriter.planFileGroups(partitionTasks);
- List<List<PositionDeletesScanTask>> groups =
ImmutableList.copyOf(plannedFileGroups);
- if (groups.size() > 0) {
- fileGroupsByPartition.put(partition, groups);
- }
- });
+ StructLikeMap<List<PositionDeletesScanTask>> filesPerPartition =
+ filesByPartition(partitionType, tasks);
- return fileGroupsByPartition;
+ return fileGroupsByPartition(partitionType, filesPerPartition);
} finally {
try {
- scanTasks.close();
+ tasks.close();
} catch (IOException io) {
LOG.error("Cannot properly close file iterable while planning for
rewrite", io);
}
}
}
- @VisibleForTesting
- RewritePositionDeletesGroup rewriteDeleteFiles(
+ private StructLikeMap<List<PositionDeletesScanTask>> filesByPartition(
+ StructType partitionType, Iterable<PositionDeletesScanTask> tasks) {
+ StructLikeMap<List<PositionDeletesScanTask>> filesByPartition =
+ StructLikeMap.create(partitionType);
+
+ for (PositionDeletesScanTask task : tasks) {
+ StructLike coerced = coercePartition(task, partitionType);
+
+ List<PositionDeletesScanTask> partitionTasks =
filesByPartition.get(coerced);
+ if (partitionTasks == null) {
+ partitionTasks = Lists.newArrayList();
+ }
+ partitionTasks.add(task);
+ filesByPartition.put(coerced, partitionTasks);
+ }
+
+ return filesByPartition;
+ }
+
+ private StructLikeMap<List<List<PositionDeletesScanTask>>>
fileGroupsByPartition(
+ StructType partitionType, StructLikeMap<List<PositionDeletesScanTask>>
filesByPartition) {
+ StructLikeMap<List<List<PositionDeletesScanTask>>> fileGroupsByPartition =
+ StructLikeMap.create(partitionType);
+
+ filesByPartition.forEach(
Review Comment:
I like this idea, implemented it. I was wondering initially as its not the
pattern of the static method like CloseableIterable.transform(), but I think
putting on instance level makes the code shorter and even more functional and
composable.
##########
spark/v3.4/spark/src/main/java/org/apache/iceberg/spark/actions/RewritePositionDeleteFilesSparkAction.java:
##########
@@ -298,7 +311,7 @@ private Result doExecuteWithPartialProgress(
.run(fileGroup -> commitService.offer(rewriteDeleteFiles(ctx,
fileGroup)));
rewriteService.shutdown();
- // Stop Commit service
+ // stop commit service
Review Comment:
Done
--
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]
