在 Google DataFusion 的 bigquery 架构中使用 arguments

Question

I am trying to use runtime arguments in bigquery schema defintion in Bigquery sink plugin. It is just two columns. Definition in argument setter.json. -

{
  "arguments" : [
   {"name":"bq.config.table","value":"activity_category"},
   
   {
   "name" : "sqloutput_schema",
   "type" : "schema",
   "value" : 
    [
    {"name":"activity_category_id","type":"string","nullable":true},
    {"name":"activity_category_description","type":"string"}
    ]
    }
   
  ]
}

Issue is in the 'sqloutput_schema', which is failing during runtime - PFA screenshot of plugin:-

Error received - Spark program 'phase-2' failed with error: Argument 'sqloutput_schema' is not defined.Please check the system logs for more details. io.cdap.cdap.api.macro.InvalidMacroException: Argument 'sqloutput_schema' is not defined. Spark program 'phase-2' failed with error: Argument 'sqloutput_schema' is not defined.Please check the system logs for more details. io.cdap.cdap.api.macro.InvalidMacroException: Argument 'sqloutput_schema' is not defined.

I am unable to find a solution as to why this is failing.

Answer 1

The problem is your schema definition. I had the same use-case and my argument was of type string and the value had the following format -

"{\"name\":\"etlSchemaBody\",\"type\":\"record\",\"fields\":
[
{\"name\":\"Id\",\"type\":\"int\"},
{\"name\":\"name\",\"type\":\"string\"}
]}"

So change the type of the argument of the schema and the schema json following the format above.