[
https://issues.apache.org/jira/browse/ASTERIXDB-1580?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15431377#comment-15431377
]
Yingyi Bu commented on ASTERIXDB-1580:
--------------------------------------
Vignesh,
In your query:
{noformat}
select a.ca_state state, count(a) cnt
from customer_address a
,customer c
,store_sales s
,date_dim d
,item i
where a.ca_address_sk = c.c_current_addr_sk
and c.c_customer_sk = s.ss_customer_sk
and s.ss_sold_date_sk = d.d_date_sk
and s.ss_item_sk = i.i_item_sk
group by a.ca_state group as stategrp
having coll_count((select value a from stategrp)) >= 10
order by cnt
limit 100;
{noformat}
After the group by, the subquery
{noformat}
(select value a from stategrp)
{noformat}
is a syntatic sugar of
{noformat}
(select value (select value g2.a from stategrp g2) from stategrp g1)
{noformat}
Basically, after grouping, "a" is a syntatic sugar of "(select value g.a from
stategrp g)", except a is referred inside a SQL-92 aggregate function symbol.
> Subquery in having clause throws ArrayIndexOutOfBoundsException
> ---------------------------------------------------------------
>
> Key: ASTERIXDB-1580
> URL: https://issues.apache.org/jira/browse/ASTERIXDB-1580
> Project: Apache AsterixDB
> Issue Type: Bug
> Reporter: Vignesh Raghunathan
> Assignee: Yingyi Bu
> Attachments: logs.zip
>
>
> The following sqlpp query throws an ArrayIndexOutOfBoundsException
> {code}
> use tpcds;
> select a.ca_state state, count(a) cnt
> from customer_address a
> ,customer c
> ,store_sales s
> ,date_dim d
> ,item i
> where a.ca_address_sk = c.c_current_addr_sk
> and c.c_customer_sk = s.ss_customer_sk
> and s.ss_sold_date_sk = d.d_date_sk
> and s.ss_item_sk = i.i_item_sk
> group by a.ca_state group as stategrp
> having coll_count((select value a from stategrp)) >= 10
> order by cnt
> limit 100;
> {code}
> tpcdsq6.zip contains the schema and data for the datasets used in the query
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
