[
https://issues.apache.org/jira/browse/CALCITE-7045?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17955010#comment-17955010
]
Mihai Budiu edited comment on CALCITE-7045 at 5/29/25 4:37 PM:
---------------------------------------------------------------
there are two possible approaches:
* have a generator that always generates fresh ids, e.g., using a counter
* use a visitor to detect ids in use and generate fresh ones when needed
In the first approach the compiler is stateful (remembers ids generated), in
the second it is stateless (all the information is in the plan) - the second
approach requires more work, though, to scan each plan when you need to
generate new ids.
In our compiler we use the first approach; rel nodes have similar ids I
believe. If you want the ids to be "dense" you can also implement a visitor to
"compact" all the corrids that appear in a plan in two passes, by collecting a
set of all used ids and renaming them 0, 1, 2, etc. This could be used before
printing plans, to ensure that most of the plans displayed today don't change.
was (Author: JIRAUSER295926):
there are two possible approaches:
* have a generator that always generates fresh ids, e.g., using a counter
* use a visitor to detect ids in use and generate fresh ones when needed
In our compiler we use the first approach; rel nodes have similar ids I
believe. If you want the ids to be "dense" you can also implement a visitor to
"compact" all the corrids that appear in a plan in two passes, by collecting a
set of all used ids and renaming them 0, 1, 2, etc. This could be used before
printing plans, to ensure that most of the plans displayed today don't change.
> Generate unique correlationId for each correlate node
> -----------------------------------------------------
>
> Key: CALCITE-7045
> URL: https://issues.apache.org/jira/browse/CALCITE-7045
> Project: Calcite
> Issue Type: Improvement
> Reporter: suibianwanwan
> Priority: Major
>
> As discussed in
> [https://lists.apache.org/thread/l5ls7hxmrkp6vqqmffxs4cq4dnv95x36] :
> Currently in SubQueryRemove, new Correlates are created based on the
> CorrelationId from the original RelNode. When this subQuery requires multiple
> Correlate expansions or when multiple subQueries share the same CorrelationId
> and are expanded separately, multiple Correlates with identical CorrelationId
> may be generated.
> This can cause difficulties during Decorrelate processing and lead to errors
> in some scenarios.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
