yihua commented on code in PR #12105:
URL: https://github.com/apache/hudi/pull/12105#discussion_r1815358599


##########
hudi-client/hudi-client-common/src/main/java/org/apache/hudi/metadata/HoodieBackedTableMetadataWriter.java:
##########
@@ -268,13 +268,11 @@ protected boolean 
initializeIfNeeded(HoodieTableMetaClient dataMetaClient,
       // If there is no commit on the dataset yet, use the 
SOLO_COMMIT_TIMESTAMP as the instant time for initial commit
       // Otherwise, we use the timestamp of the latest completed action.
       String initializationTime = 
dataMetaClient.getActiveTimeline().filterCompletedInstants().lastInstant().map(HoodieInstant::getTimestamp).orElse(SOLO_COMMIT_TIMESTAMP);
-
-      // Initialize partitions for the first time using data from the files on 
the file system
-      if (!initializeFromFilesystem(initializationTime, partitionsToInit, 
inflightInstantTimestamp)) {
+      initializeFromFilesystem(initializationTime, partitionsToInit, 
inflightInstantTimestamp);
+      if 
(!this.dataMetaClient.getTableConfig().isMetadataPartitionAvailable(MetadataPartitionType.FILES))
 {

Review Comment:
   Should this check all metadata partitions that should be initialized besides 
FILES?



##########
hudi-client/hudi-client-common/src/main/java/org/apache/hudi/metadata/HoodieBackedTableMetadataWriter.java:
##########
@@ -520,9 +517,11 @@ private Pair<Integer, HoodieData<HoodieRecord>> 
initializePartitionStatsIndex(Li
   }
 
   private Pair<Integer, HoodieData<HoodieRecord>> 
initializeColumnStatsPartition(Map<String, Map<String, Long>> 
partitionToFilesMap) {
+    // during initialization, we need stats for base and log files.
     HoodieData<HoodieRecord> records = 
HoodieTableMetadataUtil.convertFilesToColumnStatsRecords(
         engineContext, Collections.emptyMap(), partitionToFilesMap, 
dataMetaClient, dataWriteConfig.isMetadataColumnStatsIndexEnabled(),
-        dataWriteConfig.getColumnStatsIndexParallelism(), 
dataWriteConfig.getColumnsEnabledForColumnStatsIndex());
+        dataWriteConfig.getColumnStatsIndexParallelism(), 
dataWriteConfig.getColumnsEnabledForColumnStatsIndex(), true,

Review Comment:
   `fetchStatsForLogFiles` is always `true`.  Could we get rid of the argument?



-- 
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]

Reply via email to