Hi @ManojkMohan 

Can you identify the possible errors in larger pipelines? How could I achieve the required output without the above code? Your reply is not very understandable.

The reason behind my feature request is as follows: The source is a NoSQL database and the schema coming from it is uncertain. Any new data I get should be written into the table with the previous data overwritten if key already present. Since schema is uncertain and the behaviour required on merge keys is overwrite, I have written this pipeline.