[
https://issues.apache.org/jira/browse/CALCITE-5061?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17534476#comment-17534476
]
Thomas Rebele edited comment on CALCITE-5061 at 5/11/22 8:55 AM:
-----------------------------------------------------------------
The problem with the RelFieldTrimmer is that it abandons trimming in some cases
if it encounters a non-trimmable node, even if it would be helpful to trim the
children. Field trimming can be applied to UNION-all as well under certain
circumstances. However this is out of scope for the current ticket. I might
create a PR for UNION-all trimming later.
I've pushed a first draft of the alternative trimmer to
[https://github.com/thomasrebele/calcite/tree/CALCITE-5061]. I still need to
complete the implementation for Calcite (extra fields, some node types, such as
Exchange). Feel free to take a look in the meantime.
was (Author: thomas.rebele):
The problem with the RelFieldTrimmer is that it abandons trimming in some cases
if it encounters a non-trimmable node, even if it would be helpful to trim the
children. Field trimming can be applied to UNION-all as well under certain
circumstances. However this is out of scope for the current ticket. I might
create a PR for UNION-all trimming later.
I've pushed a first draft of the alternative trimmer to
[https://github.com/thomasrebele/calcite/tree/CALCITE-5061.] I still need to
complete the implementation for Calcite (extra fields, some node types, such as
Exchange). Feel free to take a look in the meantime.
> Improve recursive application of the field trimming
> ---------------------------------------------------
>
> Key: CALCITE-5061
> URL: https://issues.apache.org/jira/browse/CALCITE-5061
> Project: Calcite
> Issue Type: Improvement
> Reporter: Thomas Rebele
> Assignee: Thomas Rebele
> Priority: Major
> Fix For: 1.31.0
>
>
> The RelFieldTrimmer has some shortcomings:
> * If the plan contains certain set ops (e.g., UNION(all=false)), even if the
> operator needs all fields for the correct result (CALCITE-3399), it may still
> make sense to apply the trimming to the children. See CALCITE-5051 for an
> example.
> * Same applies for a Sort with dynamic parameters in the fetch/offset, and
> RepeatUnions
> * The makeZeroLiteral logic in trimChildRestore(...) does not work for ARRAY
> / Java types.
--
This message was sent by Atlassian Jira
(v8.20.7#820007)
