[Feature][Connector-V2] Fix Unsupported data type exception when ck inserting JSON via JDBC

[Cyanty]

close #9857

Issue:
When using clickhouse-client:0.3.2-patch11, inserting data into a JSON column throws an Unsupported data type exception. The error originates from ClickHouseRowBinaryProcessor#MappedFunctions.serialize(484).

Root Cause:
The ClickHouseRowBinaryProcessor in this driver version lacks a serializer mapping for the JSON data type. For INSERT ... VALUES (?,?,...) statements, the driver defaults to using InputBasedPreparedStatement, which invokes the serialize method during addBatch(). This inevitably triggers the exception when a JSON type is encountered.

Solution:
It was observed that if the VALUES() clause contains characters other than ?, ,, or Whitespace, the driver defaults to SqlBasedPreparedStatement. This implementation builds the INSERT statement by concatenating values as a single string and does not call the serialize method. By slightly modifying the statement, we can force the use of SqlBasedPreparedStatement, thus bypassing the bug and enabling JSON insertion without a driver update.

Check list

• If any new Jar binary package adding in your PR, please add License Notice according
  New License Guide
• If necessary, please update the documentation to describe the new feature. https://github.com/apache/seatunnel/tree/dev/docs
• If you are contributing the connector code, please check that the following files are updated:
  1. Update plugin-mapping.properties and add new connector information in it
  2. Update the pom file of seatunnel-dist
  3. Add ci label in label-scope-conf
  4. Add e2e testcase in seatunnel-e2e
  5. Update connector plugin_config

[Hisoka-X]

We already has https://github.com/apache/seatunnel/blob/dev/seatunnel-api/src/main/java/org/apache/seatunnel/api/table/type/CommonOptions.java#L33. So I think we can reuse this option. Not create new type.

[Cyanty]

We already has https://github.com/apache/seatunnel/blob/dev/seatunnel-api/src/main/java/org/apache/seatunnel/api/table/type/CommonOptions.java#L33. So I think we can reuse this option. Not create new type.

Hi @Hisoka-X, thank you for the review.
I see now that BasicType#JSON_TYPE is redundant. ck sink can infer the type from the catalog. I'll remove it in the next commit to allow writing to JSON columns without any config changes.

[zhangshenghang]

Please also verify in e2e code

[Cyanty]

Please also verify in e2e code

Thank you for the review, I'll add the e2e test in the next commit.

[Hisoka-X]

cc @JeremyXin

[Hisoka-X]

The sqltype is come from SeaTunnel. Why it can used in clickhouse sql directly? We should use ClickhouseTypeConverter to convert.

[Cyanty]

The sqltype is come from SeaTunnel. Why it can used in clickhouse sql directly? We should use ClickhouseTypeConverter to convert.

Hi @Hisoka-X , Thank you for your guidance. I'm facing a roadblock and would appreciate your input.
To force SqlBasedPreparedStatement for writing JSON types, I'm adding a CAST in the VALUES() clause. And that all other types are converted to a generic ?.
Since only the JSON type requires special handling, I'd like to skip implementing ClickhouseTypeConverter#convert and instead directly use the data types from tableSchema for the CAST.
e.g.​​,String typeName = tableSchema.getOrDefault(fieldName, ""), Do you think this is feasible?

[Hisoka-X]

I'm not sure what do you want to do. Could you share some demo? Thanks.

[Cyanty]

I'm not sure what do you want to do. Could you share some demo? Thanks.

Hi @Hisoka-X , Thank you for your reply.
Use the native ClickHouse tableSchema for type mapping in the INSERT sql, instead of using SeaTunnel types. JSON columns are now explicitly cast with CAST(? AS String), all other column types remain as a ? placeholder.
here is the code change:

// Uses ClickHouse table schema to get type names
String[] typeNames =
    Arrays.stream(fieldNames)
          .map(name -> tableSchema.getOrDefault(name, ""))
          .toArray(String[]::new);

The CAST for JSON is a workaround for older clickhouse-jdbc limitations when JSON support was experimental, then this implementation is somewhat informal, and this might require further discussion.

[Hisoka-X]

Hi @Cyanty , I found the same issue in ClickHouse/clickhouse-java#2491. I think the better way is upgrade the clickhouse-jdbc version. Could you try this way? Thanks.

[Cyanty]

Hi @Cyanty , I found the same issue in ClickHouse/clickhouse-java#2491. I think the better way is upgrade the clickhouse-jdbc version. Could you try this way? Thanks.

Hi @Hisoka-X , Of course, upgrading the driver version would be more reasonable. I will attempt to upgrade the version and check if there are any incompatible codes.

[Hisoka-X]

Thanks @Cyanty !