[ 
https://issues.apache.org/jira/browse/IGNITE-26137?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Pavel Pereslegin reassigned IGNITE-26137:
-----------------------------------------

    Assignee: Pavel Pereslegin

> Sql. Support MERGE operation in optimized plans
> -----------------------------------------------
>
>                 Key: IGNITE-26137
>                 URL: https://issues.apache.org/jira/browse/IGNITE-26137
>             Project: Ignite
>          Issue Type: Improvement
>          Components: sql ai3
>            Reporter: Konstantin Orlov
>            Assignee: Pavel Pereslegin
>            Priority: Major
>              Labels: ignite-3
>
> As of now, only single-tuple INSERT queries out of all possible DML is mapped 
> to an optimized KV-plans. Let's extend support for MERGE operation with 
> UPSERT semantic.
> The MERGE has an UPSERT semantic iff all of the listed below are true:
> * {{ON}} condition represents predicate over primary key only
> * Source specification doesn't involve any dynamic data sources (like tables, 
> system views, table functions, etc)
> * {{WHEN MATCHED THEN UPDATE}} clause explicitly enumerates updates for all 
> columns (except for pk columns) from the destination table
> * {{WHEN NOT MATCHED THEN INSERT}} clause explicitly enumerates updates for 
> all column from the destination table. Default values are not allowed
> * {{INSERT}} clause and {{UPDATE}} clause must refer to the same column of 
> source for the same target column. For example, if source clause produces 
> tuple {{(a, b, c)}}, and {{INSERT}} tuple specified as {{(a, c, b)}}, then 
> the only possible option for the {{UPDATE}} clause is to specify {{SET col_0 
> = a, col_1 = c, col_2 = b}}.
> See {{org.apache.ignite.internal.sql.engine.prepare.KeyValueModifyPlan}} for 
> reference.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

Reply via email to