[
https://issues.apache.org/jira/browse/IOTDB-339?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17004379#comment-17004379
]
Jialin Qiao commented on IOTDB-339:
-----------------------------------
[https://github.com/apache/incubator-iotdb/pull/677]
> Optimize QueryPlan
> ------------------
>
> Key: IOTDB-339
> URL: https://issues.apache.org/jira/browse/IOTDB-339
> Project: Apache IoTDB
> Issue Type: Improvement
> Reporter: Jialin Qiao
> Priority: Minor
>
> For a query such as "select s1, s1 from root.d1", the query process is as
> follows:
> (1) Parse the SQL to a QueryPlan, in which the selected paths are
> <root.d1.s1, root.d1.s1>.
> (2) Return the selected paths to the client for result printing, and
> recording this QueryPlan in the server.
> (3) When the client fetches the result, the QueryPlan is passed to
> QueryProcessExecutor. The executor deduplicate the paths and execute the plan.
> (4) Return the result set.
>
> I recommend moving the deduplication to step (1) and put the origin projected
> paths and deduplicated paths in QueryPlan. Then, the executor could focus on
> executing, instead of optimization.
>
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
