[jira] [Created] (IMPALA-9432) An "is not distinct from constant" predicate should be converted to = in the expression rewrites

["Wenzhe Zhou (Jira)" <ji...@apache.org>]

Wenzhe Zhou created IMPALA-9432:
-----------------------------------

             Summary: An "is not distinct from constant" predicate should be converted to = in the expression rewrites
                 Key: IMPALA-9432
                 URL: https://issues.apache.org/jira/browse/IMPALA-9432
             Project: IMPALA
          Issue Type: Improvement
          Components: Frontend, Planner
    Affects Versions: Impala 3.3.0
            Reporter: Wenzhe Zhou


In following query, "l_orderkey is not distinct from 1" should be converted as "l_orderkey = 1" in expression rewrite.

select count(*) from tpch_kudu.lineitem where l_orderkey is not distinct from 1 and l_partkey = 1 and l_suppkey = 1;

Here is the current plan.

---- PLAN
PLAN-ROOT SINK
|
01:AGGREGATE [FINALIZE]
| output: count(*)
| row-size=8B cardinality=1
|
00:SCAN KUDU [tpch_kudu.lineitem]
 predicates: l_orderkey IS NOT DISTINCT FROM 1
 kudu predicates: l_partkey = 1, l_suppkey = 1
 row-size=8B cardinality=1
---- SCANRANGELOCATIONS
NODE 0:
 ScanToken\{table-name=impala::tpch_kudu.lineitem, hash-partition-buckets: [0]}
 ScanToken\{table-name=impala::tpch_kudu.lineitem, hash-partition-buckets: [1]}
 ScanToken\{table-name=impala::tpch_kudu.lineitem, hash-partition-buckets: [2]}
 ScanToken\{table-name=impala::tpch_kudu.lineitem, hash-partition-buckets: [3]}
 ScanToken\{table-name=impala::tpch_kudu.lineitem, hash-partition-buckets: [4]}
 ScanToken\{table-name=impala::tpch_kudu.lineitem, hash-partition-buckets: [5]}
 ScanToken\{table-name=impala::tpch_kudu.lineitem, hash-partition-buckets: [6]}
 ScanToken\{table-name=impala::tpch_kudu.lineitem, hash-partition-buckets: [7]}
 ScanToken\{table-name=impala::tpch_kudu.lineitem, hash-partition-buckets: [8]}
---- DISTRIBUTEDPLAN
PLAN-ROOT SINK
|
03:AGGREGATE [FINALIZE]
| output: count:merge(*)
| row-size=8B cardinality=1
|
02:EXCHANGE [UNPARTITIONED]
|
01:AGGREGATE
| output: count(*)
| row-size=8B cardinality=1
|
00:SCAN KUDU [tpch_kudu.lineitem]
 predicates: l_orderkey IS NOT DISTINCT FROM 1
 kudu predicates: l_partkey = 1, l_suppkey = 1
 row-size=8B cardinality=1



--
This message was sent by Atlassian Jira
(v8.3.4#803005)