17.19. EXPLAIN ANALYZE

Synopsis

EXPLAIN ANALYZE statement

Description

Execute the statement and show the distributed execution plan of the statement along with the cost of each operation.

Note

The stats may not be entirely accurate, especially for queries that complete quickly.

Examples

In the example below, you can see the CPU time spent in each stage, as well as the relative cost of each operator in the stage. Note that the relative cost of the operators is based on wall time, which may or may not be correlated to CPU time. For each plan node you can see statistics for individual operators. Such statistics are useful when one wants to detect data anomalies for a query (skewness, abnormal hash collisions).

presto:sf1> EXPLAIN ANALYZE SELECT count(*), clerk FROM orders WHERE orderdate > date '1995-01-01' GROUP BY clerk;

                                          Query Plan
-----------------------------------------------------------------------------------------------
Fragment 1 [SINGLE]
    Cost: CPU 7.07ms, Input: 1000 rows (8.79kB), Output: 1000 rows (8.79kB)
    Output layout: [count]
    Output partitioning: SINGLE []
    - Output[Query Plan] => [count:bigint]
            Cost: 0.00%, Output: 1000 rows (8.79kB)
            TaskOutputOperator := Drivers: 1, Input avg.: 1000.00 lines, Input std.dev.: 0.00%
            Query Plan := count
        - RemoteSource[2] => [count:bigint]
                Cost: 100.00%, Output: 1000 rows (8.79kB)
                ExchangeOperator := Drivers: 1, Input avg.: 1000.00 lines, Input std.dev.: 0.00%

Fragment 2 [HASH]
    Cost: CPU 22.55ms, Input: 4000 rows (148.44kB), Output: 1000 rows (8.79kB)
    Output layout: [count]
    Output partitioning: SINGLE []
    - Project[] => [count:bigint]
            Cost: 4.11%, Input: 1000 rows (37.11kB), Output: 1000 rows (8.79kB), Filtered: 0.00%
            TaskOutputOperator := Drivers: 1, Input avg.: 1000.00 lines, Input std.dev.: 0.00%
            FilterAndProjectOperator := Drivers: 1, Input avg.: 1000.00 lines, Input std.dev.: 0.00%
        - Aggregate(FINAL)[clerk] => [clerk:varchar(15), $hashvalue:bigint, count:bigint]
                Cost: 23.29%, Output: 1000 rows (37.11kB)
                HashAggregationOperator := Drivers: 1, Input avg.: 4000.00 lines, Input std.dev.: 0.00%
                HashAggregationOperator := Collisions avg.: 112.00 (345.16% est.), Collisions std.dev.: 0.00%
                count := "count"("count_8")
            - RemoteSource[3] => [clerk:varchar(15), count_8:bigint, $hashvalue:bigint]
                    Cost: 72.60%, Output: 4000 rows (148.44kB)
                    ExchangeOperator := Drivers: 1, Input avg.: 4000.00 lines, Input std.dev.: 0.00%

Fragment 3 [tpch:orders:1500000]
    Cost: CPU 20.58s, Input: 818058 rows (22.62MB), Output: 4000 rows (148.44kB)
    Output layout: [clerk, count_8, $hashvalue_9]
    Output partitioning: HASH [clerk]
    - Aggregate(PARTIAL)[clerk] => [clerk:varchar(15), $hashvalue_9:bigint, count_8:bigint]
            Cost: 5.37%, Output: 4000 rows (148.44kB)
            HashAggregationOperator := Drivers: 4, Input avg.: 204514.50 lines, Input std.dev.: 0.05%
            HashAggregationOperator := Collisions avg.: 5701.28 (17569.93% est.), Collisions std.dev.: 1.12%
            PartitionedOutputOperator := Drivers: 4, Input avg.: 1000.00 lines, Input std.dev.: 0.00%
            count_8 := "count"(*)
        - ScanFilterProject[table = tpch:tpch:orders:sf1.0, originalConstraint = ("orderdate" > "$literal$date"(BIGINT '9131')), filterPredicate = ("orderdate" > "$literal$date"(BIGINT '9131'))] => [cler
                Cost: 94.63%, Input: 1500000 rows (0B), Output: 818058 rows (22.62MB), Filtered: 45.46%
                ScanFilterAndProjectOperator := Drivers: 4, Input avg.: 375000.00 lines, Input std.dev.: 0.00%
                $hashvalue_9 := "combine_hash"(BIGINT '0', COALESCE("$operator$hash_code"("clerk"), 0))
                orderdate := tpch:orderdate
                clerk := tpch:clerk

See Also

EXPLAIN