[
https://issues.apache.org/jira/browse/PHOENIX-5451?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16912120#comment-16912120
]
chenglei edited comment on PHOENIX-5451 at 8/21/19 9:27 AM:
------------------------------------------------------------
Uploaded my first patch, this patch is mainly for:
# For Join, added a new {{JoinTable.pruneSubselectAliasedNodes}} at the end of
{{JoinCompiler.compile}} to pruning columns for each {{JoinCompiler.Table}}
according to {{JoinCompiler.columnNodes}}.
# For subquery, added a new
{{SubselectRewriter.pruneInnerSubselectAliasedNodes}} at the end of
{{SubselectRewriter.flatten}} to pruning columns for the
{{DerivedTableNode.getSelect}} according to
{{ParseNodeUtil.collectReferencedColumnNamesForSingleTable}}.
[~tdsilva], please help me have a review, the PR is
https://github.com/apache/phoenix/pull/568
was (Author: comnetwork):
Uploaded my first patch, this patch is mainly for:
#For Join, added a new {{JoinTable.pruneSubselectAliasedNodes}} at the end of
{{JoinCompiler.compile}} to pruning columns for each {{JoinCompiler.Table}}
according to {{JoinCompiler.columnNodes}}.
#For subquery, added a new
{{SubselectRewriter.pruneInnerSubselectAliasedNodes}} at the end of
{{SubselectRewriter.flatten}} to pruning columns for the
{{DerivedTableNode.getSelect}} according to
{{ParseNodeUtil.collectReferencedColumnNamesForSingleTable}}.
[~tdsilva], please help me have a review, the PR is
https://github.com/apache/phoenix/pull/568
> Column pruning for Subselect
> ----------------------------
>
> Key: PHOENIX-5451
> URL: https://issues.apache.org/jira/browse/PHOENIX-5451
> Project: Phoenix
> Issue Type: Improvement
> Affects Versions: 4.14.2
> Reporter: chenglei
> Assignee: chenglei
> Priority: Major
> Fix For: 4.15.0, 5.1.0
>
> Attachments: PHOENIX-5451_v1-4.x-HBase-1.4.patch
>
> Time Spent: 10m
> Remaining Estimate: 0h
>
> Phoenix has already could prune column for single flat table query, but for
> subselect in subquery or join, column pruning is not implemented.
> Consider following join sql :
> {code:java}
> select a.aid,b.bid from
> (select aid,age from table1 where age >=11 and age<=33 order by age limit 3)
> a inner join
> (select bid,code from table2 order by code limit 1) b on a.aid=b.bid
> {code}
> for LHS, {{table1.age}} is not referenced by outer join, and for RHS,
> {{table2.code}} is also not referenced by outer join, {{table1.age}} and
> {{table2.code}} could both be pruned when execute query for LHS and RHS, but
> now Phoenix does not prune them.
> Another example is subquery sql:
> {code:java}
> SELECT t.eid id, t.bstr b FROM
> (SELECT entity_id eid, a_string astr, b_string bstr FROM table WHERE a_byte
> + 1 < 9 limit 2) t ORDER BY b,id
> {code}
> for inner subselect sql, {{a_string}} is never used by outer sql, so
> {{a_string}} could be pruned when execute query for inner subselect, but now
> Phoenix does not prune it.
> Some existing test cases could be optimized by this patch.
--
This message was sent by Atlassian Jira
(v8.3.2#803003)
