sankarh commented on a change in pull request #579: HIVE-21109 : Support stats replication for ACID tables. URL: https://github.com/apache/hive/pull/579#discussion_r269880083
########## File path: ql/src/java/org/apache/hadoop/hive/ql/metadata/Hive.java ########## @@ -2950,21 +2956,33 @@ public Partition createPartition(Table tbl, Map<String, String> partSpec) throws int size = addPartitionDesc.getPartitionCount(); List<org.apache.hadoop.hive.metastore.api.Partition> in = new ArrayList<org.apache.hadoop.hive.metastore.api.Partition>(size); - AcidUtils.TableSnapshot tableSnapshot = AcidUtils.getTableSnapshot(conf, tbl, true); long writeId; String validWriteIdList; - if (tableSnapshot != null && tableSnapshot.getWriteId() > 0) { - writeId = tableSnapshot.getWriteId(); - validWriteIdList = tableSnapshot.getValidWriteIdList(); + + // In case of replication, get the writeId from the source and use valid write Id list + // for replication. + if (addPartitionDesc.getReplicationSpec() != null && + addPartitionDesc.getReplicationSpec().isInReplicationScope() && + addPartitionDesc.getPartition(0).getWriteId() > 0) { + writeId = addPartitionDesc.getPartition(0).getWriteId(); + validWriteIdList = Review comment: Even that logic to create ValidWriteIdList based on all repl opened txns isn't right as it says that stats are valid for all these open txns but it isn't. Also, it sets high water mark based on 0th index in the replTxnsList map which might be pointing to wrong writeId compared to current txn's writeId. So, I doubt, this logic should be anyways removed. ---------------------------------------------------------------- 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. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services