Check query execution

You can determine if query vectorization occurred during execution by running the EXPLAIN VECTORIZATION query statement.

  1. Start Hive from Beeline.
    $ hive
  2. Set hive.explain.user to false to see vector values.
    SET hive.explain.user=false;
  3. Run the EXPLAIN VECTORIZATION statement on the query you want CDP to process using vectorization.
    EXPLAIN VECTORIZATION SELECT COUNT(*) FROM employees where emp_no>10;
    The following output confirms that vectorization occurs:
    +----------------------------------------------------+
    |                      Explain                       |
    +----------------------------------------------------+
    | Plan optimized by CBO.                             |
    |                                                    |
    | Vertex dependency in root stage                    |
    | Reducer 2 <- Map 1 [CUSTOM_SIMPLE_EDGE] *vectorized* |
    |       File Output Operator [FS_14]                 |
    |         Group By Operator [GBY_13] (rows=1 width=12) |
    |           Output:["_col0"],aggregations:["count(VALUE._col0)"] |
    |         <-Map 1 [CUSTOM_SIMPLE_EDGE] vectorized    | 
    |           PARTITION_ONLY_SHUFFLE [RS_12]           |
    |             Group By Operator [GBY_11] (rows=1 width=12) |
    |               Output:["_col0"],aggregations:["count()"] |
    |               Select Operator [SEL_10] (rows=1 width=4) |
    |                 Filter Operator [FIL_9] (rows=1 width=4) |
    |                   predicate:(emp_no > 10)          |
    |                   TableScan [TS_0] (rows=1 width=4) |
    |                     default@employees,employees,Tbl:COMPLETE,Col:NONE,Output:["emp_no"] |
    |                                                    |
    +----------------------------------------------------+
    23 rows selected (1.542 seconds)