[
https://issues.apache.org/jira/browse/PHOENIX-2679?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Maryann Xue updated PHOENIX-2679:
---------------------------------
Attachment: suggested-calcite-changes-for-PHOENIX-2679.patch
PHOENIX-2679.wip.2.patch
That would be good, [~julianhyde]. We are fine for now though.
I've made most of the test cases work in CalciteXXXIT now. Also tested with
MultiCFQueryExecIT. There were a couple of failures due to other issues.
I am attaching a patch for the Phoenix side as well as a "suggested changes"
patch for Calcite side. The suggested changes are:
1. Change the logic of "Prepare.shouldTrim()". I think the original logic is
wrong, and trimming on structured types after flattening did not work correctly.
2. Call "SqlToRelConverter.flattenTypes()" in addition to
"Prepare.flattenTypes()" so that the real flattening will be done.
3. Change behavior of Hook.TRIMMED, in order to allow Phoenix to do some
customized handling (i.e., shadow table replacement) before "optimize()" is
called. Because otherwise the "root" before replacement would be registered as
the "original root" in the planner, which would create some mess afterwards,
like trying to find applicable materializations using the original root, and
taking the original root into planning space.
4. Enhancement to RelStructuredTypeFlattener: add RexSubQuery handling, correct
RelCorrelVariable handling.
I will open a JIRA in Calcite regarding (4). [~julianhyde], any suggestions on
(1), (2) and (3)?
> Implement column family schema structure in Calcite-Phoenix
> -----------------------------------------------------------
>
> Key: PHOENIX-2679
> URL: https://issues.apache.org/jira/browse/PHOENIX-2679
> Project: Phoenix
> Issue Type: Task
> Reporter: Maryann Xue
> Assignee: Maryann Xue
> Labels: calcite
> Attachments: PHOENIX-2679.wip.2.patch, PHOENIX-2679.wip.patch,
> suggested-calcite-changes-for-PHOENIX-2679.patch
>
>
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
