Can we pass partition key as a parameter in a generic/dynamic azure dataflow.

Question

Operation on target Load Entity failed: Operation on target Load Entity failed: {"StatusCode":"DFExecutorUserError","Message":"Job failed due to reason: com.microsoft.dataflow.Issues: DF-TX-114 - Key partitioning does not allow computed columns - [564 737],\nsource(\n) ~> DLExtract,EXE-0001,Dataflow cannot be analyzed as a graph\nDF-TX-114 - Key partitioning does not allow computed columns - [564 737],EXE-0001,Dataflow cannot be analyzed as a graph,\nsource(\n) ~> DBExtract\nDF-TX-114 - Key partitioning does not allow computed columns - EXE-0001,Dataflow cannot be analyzed as a graph,\nInsertRows sink(\n) ~> Insert2DB,[564 737]\nDF-TX-114 - Key partitioning does not allow computed columns - \nMarkasUpdate sink(\n) ~> Update2DB,EXE-0001,Dataflow cannot be analyzed as a graph,[564 737]","Details":""}

Answer 1

Hi Batta, Uha Kanakavalli

There is a limitation in Key Paritioning in data flow. It doesnt allow dynamic expressions in key field.

If you want to use such dynamic expressions, you can use a derived column and use that in the paritition key field. But not directly on the key field.

Below thread discusses a similar issue where we use data flow parameter for key column.

https://stackoverflow.com/questions/75225600/azure-data-factory-dataflow-error-key-partitioning-does-not-allow-computed-colu

Below are my test data.

If i try to use month(toDate(date)) in the partition key column directly, i am getting same error as yours. So I add a derived column and then use the expression there as follows.

Please check and let us know for any question.

Thanks