sivabalan narayanan created HUDI-4181:
-----------------------------------------
Summary: Async indexer fails for simple index creation
Key: HUDI-4181
URL: https://issues.apache.org/jira/browse/HUDI-4181
Project: Apache Hudi
Issue Type: Bug
Components: metadata
Reporter: sivabalan narayanan
I tried to ran async indexing to build out col stats and ran into issues. Had
to make following fix to make progress. but this needs proper fix.
Local fix that worked for me.
{code:java}
diff --git
a/hudi-client/hudi-client-common/src/main/java/org/apache/hudi/metadata/HoodieBackedTableMetadataWriter.java
b/hudi-client/hudi-client-common/src/main/java/org/apache/hudi/metadata/HoodieBackedTableMetadataWriter.java
index f5a96fb676..1e67020810 100644
---
a/hudi-client/hudi-client-common/src/main/java/org/apache/hudi/metadata/HoodieBackedTableMetadataWriter.java
+++
b/hudi-client/hudi-client-common/src/main/java/org/apache/hudi/metadata/HoodieBackedTableMetadataWriter.java
@@ -955,7 +955,6 @@ public abstract class HoodieBackedTableMetadataWriter
implements HoodieTableMeta
HoodieTableFileSystemView fsView =
HoodieTableMetadataUtil.getFileSystemView(metadataMetaClient);
for (Map.Entry<MetadataPartitionType, HoodieData<HoodieRecord>> entry :
partitionRecordsMap.entrySet()) {
final String partitionName = entry.getKey().getPartitionPath();
- final int fileGroupCount = entry.getKey().getFileGroupCount();
HoodieData<HoodieRecord> records = entry.getValue();
List<FileSlice> fileSlices =
@@ -965,9 +964,10 @@ public abstract class HoodieBackedTableMetadataWriter
implements HoodieTableMeta
// so if there are no committed file slices, look for inflight slices
fileSlices =
HoodieTableMetadataUtil.getPartitionLatestFileSlicesIncludingInflight(metadataMetaClient,
Option.ofNullable(fsView), partitionName);
}
- ValidationUtils.checkArgument(fileSlices.size() == fileGroupCount,
+ final int fileGroupCount = fileSlices.size();
+ /*ValidationUtils.checkArgument(fileSlices.size() == fileGroupCount,
String.format("Invalid number of file groups for partition:%s,
found=%d, required=%d",
- partitionName, fileSlices.size(), fileGroupCount));
+ partitionName, fileSlices.size(), fileGroupCount));*/
List<FileSlice> finalFileSlices = fileSlices;
HoodieData<HoodieRecord> rddSinglePartitionRecords = records.map(r -> {
{code}
--
This message was sent by Atlassian Jira
(v8.20.7#820007)