bryanck commented on code in PR #5143:
URL: https://github.com/apache/iceberg/pull/5143#discussion_r908562682
##########
spark/v3.2/spark/src/main/java/org/apache/iceberg/spark/source/SparkBatch.java:
##########
@@ -65,18 +63,24 @@ public InputPartition[] planInputPartitions() {
Broadcast<Table> tableBroadcast =
sparkContext.broadcast(SerializableTable.copyOf(table));
String expectedSchemaString = SchemaParser.toJson(expectedSchema);
- InputPartition[] readTasks = new InputPartition[tasks.size()];
+ InputPartition[] readTasks = new InputPartition[tasks().size()];
Tasks.range(readTasks.length)
.stopOnFailure()
.executeWith(localityEnabled ? ThreadPools.getWorkerPool() : null)
.run(index -> readTasks[index] = new ReadTask(
- tasks.get(index), tableBroadcast, expectedSchemaString,
+ tasks().get(index), tableBroadcast, expectedSchemaString,
Review Comment:
It looked like there are other cases where concurrent access would cause
issues so I decided just to revert it back to the way it was in 0.13 in
principle. For example in `SparkBatchQueryScan` both `tasks()` and `files()`
behave similarly. So we'd need to update both if we want to support concurrency.
--
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]
