Wenhai created ASTERIXDB-1839:
---------------------------------
Summary: Subquery de-correlation problem.
Key: ASTERIXDB-1839
URL: https://issues.apache.org/jira/browse/ASTERIXDB-1839
Project: Apache AsterixDB
Issue Type: Bug
Components: AsterixDB
Environment: MAC linux
Reporter: Wenhai
Priority: Critical
When setup a subquery before an effective query and return a simple list, we
get a problem regarding the subquery decorrelation.
Query on tpch LineItem:
{noformat}
use dataverse tpch;
let $c := count(
for $d in dataset LineItem
return $d
)
for $g in [1,2,3,4,5]
return $g
{noformat}
Plan
{noformat}
distribute result [%0->$$2]
-- DISTRIBUTE_RESULT |PARTITIONED|
exchange
-- ONE_TO_ONE_EXCHANGE |PARTITIONED|
join (TRUE)
-- NESTED_LOOP |PARTITIONED|
exchange
-- ONE_TO_ONE_EXCHANGE |PARTITIONED|
project ([])
-- STREAM_PROJECT |PARTITIONED|
exchange
-- ONE_TO_ONE_EXCHANGE |PARTITIONED|
data-scan []<-[$$8, $$9, $$0] <- tpch:LineItem
-- DATASOURCE_SCAN |PARTITIONED|
exchange
-- ONE_TO_ONE_EXCHANGE |PARTITIONED|
empty-tuple-source
-- EMPTY_TUPLE_SOURCE |PARTITIONED|
exchange
-- BROADCAST_EXCHANGE |PARTITIONED|
unnest $$2 <- function-call: asterix:scan-collection,
Args:[AOrderedList: [ AInt64: {1}, AInt64: {2}, AInt64: {3}, AInt64: {4},
AInt64: {5} ]]
-- UNNEST |UNPARTITIONED|
aggregate [] <- []
-- AGGREGATE |UNPARTITIONED|
empty-tuple-source
-- EMPTY_TUPLE_SOURCE |UNPARTITIONED|
{noformat}
Is this plan correct?
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
