Can someone please help me here?. I am facing issue in modelling table as the input event is highly nested and i am not able to access the nested element from the code snippet. https://apache-flink.slack.com/archives/C03G7LJTS2G/p1684852955178819

@Shiv Desai Like mentioned in the channel description, one of the few rules we have in this Slack is not direct @ people

Sorry for that. Could you please help me as i am stuck with the above mentioned issue.

I can’t. My experience on switching from the DataStream API to the Table API is very little. Not sure why you can’t read from the Schema Registry (confluent-avro) in the Table API to avoid the conversion

I tried the below code to avoid the conversion and it is working fine for root element("a") in the json but i am not sure what datatype should i give for "b". Furthermore, how can i access "c" using select statement. Example:

{
"a":"a"
"b": {
       "c": "c"
     }
}

Code:

tEnv.executeSql("CREATE TABLE rules (\n" +
         "    a VARCHAR(3),  \n" +
         ") WITH (\n" +
         "    'connector' = 'kafka',\n" +
         "    'properties.bootstrap.servers' = '',\n" +
         "    'scan.startup.mode' = 'earliest-offset',\n" +
         "    'topic'     = '',\n" +
         "    'value.format'    = 'avro-confluent',\n" +
         "    'value.avro-confluent.url' = ''\n" +
         ")");

It would be a ROW https://nightlies.apache.org/flink/flink-docs-master/docs/dev/table/types/#row

Thank you very much for you help. Furthermore, i was curious and wanted to know whether is there a way in which we can create the Table with all fields using the Schema string directly. ?.

There isn't in open source Flink, there is https://issues.apache.org/jira/browse/FLINK-18777 but that has been stalled for quite some time