singhpk234 commented on a change in pull request #3676:
URL: https://github.com/apache/iceberg/pull/3676#discussion_r826230903
##########
File path: core/src/main/java/org/apache/iceberg/actions/BinPackStrategy.java
##########
@@ -155,9 +155,12 @@ public RewriteStrategy options(Map<String, String>
options) {
public Iterable<List<FileScanTask>> planFileGroups(Iterable<FileScanTask>
dataFiles) {
ListPacker<FileScanTask> packer = new
BinPacking.ListPacker<>(maxGroupSize, 1, false);
List<List<FileScanTask>> potentialGroups = packer.pack(dataFiles,
FileScanTask::length);
+
return potentialGroups.stream().filter(group ->
- group.size() >= minInputFiles || sizeOfInputFiles(group) >
targetFileSize ||
- group.stream().anyMatch(this::taskHasTooManyDeletes)
+ (group.size() >= minInputFiles && group.size() > 1) ||
+ sizeOfInputFiles(group) > targetFileSize ||
+ group.stream().anyMatch(this::taskHasTooManyDeletes) ||
+ (group.stream().anyMatch(this::taskHasDeletes) && group.size()
== 1)
Review comment:
I am thinking of a case like Group with (1 Data file with 1 delete) with
deleteFileThreshold = 2
3 as per my understanding would return false I beleive as deletes = 1 which
is < 2
```java
private boolean taskHasTooManyDeletes(FileScanTask task) {
return task.deletes() != null && task.deletes().size() >=
deleteFileThreshold;
}
```
4 as per my understanding would return return true as delete = 1 which >= 1
```java
private boolean taskHasDeletes(FileScanTask task) {
return task.deletes() != null && task.deletes().size() >= 1;
}
```
Appologies seems like I am missing something here, can you please elaborate
more
--
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]
