[
https://issues.apache.org/jira/browse/IMPALA-2792?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17060207#comment-17060207
]
Gabor Kaszab commented on IMPALA-2792:
--------------------------------------
Hive also support size() that is kind of identical to count(). We can implement
both to keep compatibility between the two systems.
> Syntactic sugar for computing aggregates over nested collections.
> -----------------------------------------------------------------
>
> Key: IMPALA-2792
> URL: https://issues.apache.org/jira/browse/IMPALA-2792
> Project: IMPALA
> Issue Type: New Feature
> Components: Frontend
> Affects Versions: Impala 2.3.0
> Reporter: Alexander Behm
> Priority: Major
> Labels: complextype, nested_types, planner, ramp-up, usability
>
> For user convenience and SQL brevity, we should add syntax extensions to
> concisely express aggregates over nested collections. Internally, we should
> re-write the concise versions into the more verbose equivalent with a
> correlated inline view.
> Example A:
> {code}
> New syntax:
> select count(c.orders) from customer c
> Internally rewrite to:
> select cnt from customer c, (select count(*) from c.orders) v
> {code}
> Example B:
> {code}
> New syntax:
> select avg(c.orders.items.price) from customer c
> Internally rewrite to:
> select a from customer c, (select avg(price) from c.orders.items) v
> {code}
> I suggest performing the rewrite inside StmtRewriter.java after rewriting all
> subqueries from the WHERE clause.
> Similar syntactic improvements should be considered for analytic functions on
> nested collections.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
