[
https://issues.apache.org/jira/browse/FLINK-33606?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17788838#comment-17788838
]
Sangram Reddy commented on FLINK-33606:
---------------------------------------
Hi
We will not be breaking any standard and Flink doesn't have to do any specific
implementations for different systems.
Ask here is to expose Flink specific functionality along with JDBC common spec.
This is not uncommon for other databases.
For example, see the postgres specific implementation which is not part of the
JDBC spec.
Postgres specific metadata interface:
[https://github.com/pgjdbc/pgjdbc/blob/ab1f2914a21edf9f8b7bfb35ad0d08789723b19f/pgjdbc/src/main/java/org/postgresql/PGResultSetMetaData.java]
And the implementation along with JDBC spec:
[https://github.com/pgjdbc/pgjdbc/blob/ab1f2914a21edf9f8b7bfb35ad0d08789723b19f/pgjdbc/src/main/java/org/postgresql/jdbc/PgResultSetMetaData.java|https://github.com/pgjdbc/pgjdbc/blob/ab1f2914a21edf9f8b7bfb35ad0d08789723b19f/pgjdbc/src/main/java/org/postgresql/jdbc/PgResultSetMetaData.java#L29]
For example, getBaseColumnName() is not part of JDBC spec
([https://docs.oracle.com/javase/8/docs/api/java/sql/ResultSetMetaData.html] ),
but it still exposed as it is postgres specific feature.
We will expose Flink specific metadata, whether to use it or not is upto the
clients.
> Support streaming mode in Flink JDBC driver
> -------------------------------------------
>
> Key: FLINK-33606
> URL: https://issues.apache.org/jira/browse/FLINK-33606
> Project: Flink
> Issue Type: Improvement
> Components: Table SQL / JDBC
> Affects Versions: 1.18.0
> Reporter: Sangram Reddy
> Priority: Major
>
> With Flink's 1.18 release we now have JDBC driver for Flink. This is a very
> useful addition as there are plenty of tools which can now connect to Flink
> and run SQL.
> However, one notable difference with other JDBC drivers vs. Flink is the
> streaming/dynamic tables where the data is continously updated. For example,
> we can run an aggregate query with GROUP BY and the groups are always being
> updated as the fresh data arrives.
> This is extremely useful for real-time monitoring applications. However, the
> current JDBC driver implementation only supports batch mode.
> This is Jira is to add support for streaming mode for Flink JDBC driver.
>
> Implementation details:
> FlinkResultSet doesn't have any additional methods to expose RowKind or the
> unique ID for that row. However this information is already available in
> FlinkResultSet class, it just needs to be exposed via methods.
>
> Wiki page with relevant details:
> https://cwiki.apache.org/confluence/display/FLINK/FLIP-293%3A+Introduce+Flink+Jdbc+Driver+For+Sql+Gateway
>
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
