You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@flink.apache.org by "Mao Jiayi (Jira)" <ji...@apache.org> on 2026/05/27 07:01:00 UTC

[jira] [Commented] (FLINK-39758) Fix error in canal json mysqlType when mysql type is varchar with no length

    [ https://issues.apache.org/jira/browse/FLINK-39758?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18083749#comment-18083749 ] 

Mao Jiayi commented on FLINK-39758:
-----------------------------------

I'll take this and plan to open a PR with the fix later this week.

> Fix error in canal json mysqlType when mysql type is varchar with no length
> ---------------------------------------------------------------------------
>
>                 Key: FLINK-39758
>                 URL: https://issues.apache.org/jira/browse/FLINK-39758
>             Project: Flink
>          Issue Type: Bug
>          Components: Flink CDC
>            Reporter: Mao Jiayi
>            Priority: Major
>              Labels: pull-request-available
>
> When using the Kafka pipeline connector with Canal JSON format and the *InferSchemaStrategy.MYSQL_TYPE* schema inference strategy, the job may fail when MySQL columns have missing or invalid length information.
> The failure happens in {*}MySqlTypeUtils{*}. The previous implementation does not properly validate column length before type conversion. When the MySQL binlog metadata contains columns with {*}length <= 0{*}, the code directly passes these invalid values to Flink's *DataTypes* constructors. For *VARCHAR* type, the code only checks {{== 0}} but not negative values. For *DECIMAL* type, there is no validation for non-positive length. For *VARBINARY* type, there is no length check at all.
> This issue occurs when parsing Canal JSON messages where certain MySQL column schemas have undefined lengths, causing invalid data type creation or runtime exceptions during schema inference.



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