I have a source table with ‘D’, ‘I’, and ‘U’ rows, I want to split the ‘D’ rows from ‘U’ and ‘I’ and load them separately into two different folders.

I used an Azure data flow conditional split to split these and pointed the streams to different streams and sinks respectively, Delete Sink for ‘D’ rows and InsertUpdate Sink for ‘I’ and ‘U’ rows.

My pipeline is dynamic, I haven’t imported any schema, and I don’t want to, and all the columns drift through the dataflow from source to sink. The conditional split is also done using the byName() function inside the expression builder.

My problem is, for example, if there are no ‘D’ rows, no columns will drift to the Delete sink, this throws an error stating ‘a sink cannot have 0 columns.

Kindly help me solve this issue, if possible a step-by-step explanation text/visual to resolve my problem.