bithw1

Thanks @bartosz25 . I print out the message from WholeStageCodegenExec#doExecute, ``` override def doExecute(): RDD[InternalRow] = { val (ctx, cleanedSource) = doCodeGen() println("cleanedSource#" + cleanedSource.body) //and println to print the...

thanks @bartosz25 . Yes, The `debugCodegen` feature prints exactly the same thing as I did by inserting a println statement in the code, as debugCodegen will call WholeStageCodegenExec#doCodegen method. hmm...I...

@hanbt

Thanks @bartosz25, that's huge great! I have learnt a lot from your articles, thanks a lot for your effort and knowledge!

Thanks @bartosz25 . I saw it. Many thanks to you for your great effort! I will learn and try

Thanks @bartosz25 , that's great, waiting for your great articles, :-)

> @bithw1 Why are you seeting this - option(DataSourceWriteOptions.HIVE_TABLE_SERDE_PROPERTIES.key, s"primaryKey=$primaryKey"). Remove this and try Thanks @ad1happy2go , let me try,thanks.

btw, i tried a little more, when I create the hudi table as a partitioned table, then i am able to insert from spark sql cli, the problem should be...

> The non-partitioned key generator is right, did you specify the key generator on Spark side? @danny0405 I don't think non-partitioned key generator is right here. When I am creating...

Thanks @rangareddy for the clarification and clue. Per your guide, I modify my test case by adding `hoodie.table.keygenerator.class=org.apache.hudi.keygen.ComplexAvroKeyGenerator.` to the table creation ddl (run by flink sql), but when I...