Two Bugs for Estimating Row Sizes in GroupByOperator
----------------------------------------------------

                 Key: HIVE-1737
                 URL: https://issues.apache.org/jira/browse/HIVE-1737
             Project: Hive
          Issue Type: Bug
            Reporter: Siying Dong
            Assignee: Siying Dong


Two bugs:
1. if UDAF uses string type, Group-by will break as it tries to insert an 
ArrayList<Object> to a HashMap<KeyWrapper, AggregationBuffer[]>.
2. The code to sample size of keys only handles String type and Text type, 
while in most cases, they are org.apache.hadoop.hive.serde2.lazy.LazyString, so 
that 0 is always used.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.

Reply via email to