[HIVE-10484] Vectorization : RuntimeException "Big Table Retained Mapping duplicate column"

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 1.2.0
• Fix Version/s: 1.2.0
• Component/s: Tez, Vectorization
• Assignee: Matt McCline
• Reporter: Mostafa Mokhtar
• Created: 24/Apr/15 22:29
• Updated: 18/May/15 19:56
• Resolved: 07/May/15 00:37

Description

With vectorization and tez enabled TPC-DS Q70 fails with

Caused by: java.lang.RuntimeException: Big Table Retained Mapping duplicate column 6 in ordered column map {6=(value column: 6, type name: int), 21=(value column: 21, type name: float), 22=(value column: 22, type name: int)} when adding value column 6, type int
	at org.apache.hadoop.hive.ql.exec.vector.VectorColumnOrderedMap.add(VectorColumnOrderedMap.java:97)
	at org.apache.hadoop.hive.ql.exec.vector.VectorColumnOutputMapping.add(VectorColumnOutputMapping.java:40)
	at org.apache.hadoop.hive.ql.exec.vector.mapjoin.VectorMapJoinCommonOperator.determineCommonInfo(VectorMapJoinCommonOperator.java:320)
	at org.apache.hadoop.hive.ql.exec.vector.mapjoin.VectorMapJoinCommonOperator.<init>(VectorMapJoinCommonOperator.java:254)
	at org.apache.hadoop.hive.ql.exec.vector.mapjoin.VectorMapJoinGenerateResultOperator.<init>(VectorMapJoinGenerateResultOperator.java:89)
	at org.apache.hadoop.hive.ql.exec.vector.mapjoin.VectorMapJoinInnerGenerateResultOperator.<init>(VectorMapJoinInnerGenerateResultOperator.java:97)
	at org.apache.hadoop.hive.ql.exec.vector.mapjoin.VectorMapJoinInnerLongOperator.<init>(VectorMapJoinInnerLongOperator.java:79)
	... 49 more

Query

 select s_state
               from  (select s_state as s_state, sum(ss_net_profit),
                             rank() over ( partition by s_state order by sum(ss_net_profit) desc) as ranking
                      from   store_sales, store, date_dim
                      where  d_month_seq between 1193 and 1193+11
                            and date_dim.d_date_sk = store_sales.ss_sold_date_sk
                            and store.s_store_sk  = store_sales.ss_store_sk
                      group by s_state
                     ) tmp1
               where ranking <= 5