[ 
https://issues.apache.org/jira/browse/IGNITE-27496?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Pavel Pereslegin updated IGNITE-27496:
--------------------------------------
    Description: 
Currently, if the batch of DDL commands could not be executed (due to an error 
in one of the commands), the fallback procedure occurs, and the commands are 
executed sequentially one by one.
This may be ineffective, especially if the problem is with the last command in 
the batch.

We need to modify the execution of the commands batch in such a way that it is 
possible to get the command number in the batch, the validation of which was 
unsuccessful, and then try again before the subbatch.

We have two options:
* Modify CatalogManager.execute so that it does not throw the exception 
directly, but instead writes it to CatalogApplyResult along with the number 
from the passed command list.
* Wrap CatalogValidationException into something like CatalogCOntextException 
and add required metadata to this exception.

  was:
Currently, if the batch of DDL commands could not be executed (due to an error 
in one of the commands), the fallback procedure occurs, and the commands are 
executed sequentially one by one.
This may be ineffective, especially if the problem is with the last command in 
the batch.

The fallback procedure needs to be improved to make it more effective.


> Sql. Improve the fallback mechanism of DDL batch execution
> ----------------------------------------------------------
>
>                 Key: IGNITE-27496
>                 URL: https://issues.apache.org/jira/browse/IGNITE-27496
>             Project: Ignite
>          Issue Type: Improvement
>          Components: catalog ai3, sql ai3
>            Reporter: Pavel Pereslegin
>            Priority: Major
>              Labels: ignite-3
>
> Currently, if the batch of DDL commands could not be executed (due to an 
> error in one of the commands), the fallback procedure occurs, and the 
> commands are executed sequentially one by one.
> This may be ineffective, especially if the problem is with the last command 
> in the batch.
> We need to modify the execution of the commands batch in such a way that it 
> is possible to get the command number in the batch, the validation of which 
> was unsuccessful, and then try again before the subbatch.
> We have two options:
> * Modify CatalogManager.execute so that it does not throw the exception 
> directly, but instead writes it to CatalogApplyResult along with the number 
> from the passed command list.
> * Wrap CatalogValidationException into something like CatalogCOntextException 
> and add required metadata to this exception.



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

Reply via email to