[ 
https://issues.apache.org/jira/browse/IGNITE-19407?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17730042#comment-17730042
 ] 

Pavel Pereslegin commented on IGNITE-19407:
-------------------------------------------

Command syntax:
{code:java}
ALTER TABLE <table_name> ALTER COLUMN <col_name> <action>;
<action> ::= SET DATA TYPE <new_type> [NOT NULL | NULLABLE] [DEFAULT <default 
value>]| SET NOT NULL | DROP NOT NULL | SET DEFAULT <default value> | DROP 
DEFAULT{code}
Example of changing the data type and dropping NOT NULL at the same time:
{code:java}
SET DATA TYPE BIGINT NULLABLE{code}

Example of changing the data type and dropping default at the same time:
{code:java}
SET DATA TYPE BIGINT DEFAULT NULL{code}
 


List of possible actions to change the column.
 - NOT NULL -> NULLABLE allowed for any non-PK column.
 - NULLABLE -> NOT NULL forbidden.

 - DEFAULT -> ANY LITERAL DEFAULT allowed for any column.
 - DROP DEFAULT / SET DEFAULT NULL allowed for any column.

 - TYPE SCALE change forbidden
 - a. TYPE PRECISION increase allowed for DECIMAL non PK column
 - b. TYPE LENGTH increase allowed for STRING and BYTE_ARRAY non PK column.
 - The following type transitions allowed for TYPE of non PK column:
FLOAT -> DOUBLE
INT8 -> INT16 -> INT32 -< INT64
Any other type transitions are forbidden.

> Sql. Introduce DDL command for ALTER COLUMN TYPE.
> -------------------------------------------------
>
>                 Key: IGNITE-19407
>                 URL: https://issues.apache.org/jira/browse/IGNITE-19407
>             Project: Ignite
>          Issue Type: Improvement
>          Components: sql
>            Reporter: Andrey Mashenkov
>            Assignee: Pavel Pereslegin
>            Priority: Major
>              Labels: ignite-3
>             Fix For: 3.0.0-beta2
>
>          Time Spent: 5.5h
>  Remaining Estimate: 0h
>
> Introduce DDL command class.
> Implement command validation logic. 
> Pass command to the CatalogService via DdlHandler.
> Validator should allow changing column DEFAULT and changing column type for 
> the next cases
> {noformat}
>   INT8 -> INT16 -> INT32 -> INT64
>   FLOAT -> DOUBLE
>   DECIMAL(10,5) -> DECIMAL(15, 5) // increasing precision only
>   VARCHAR(10) -> VARCHAR(20)
>   VARBINARY(10) -> VARBINARY(20)
>    NOT NULL -> NULLABLE
> {noformat}



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

Reply via email to