Can I avoiding mapping an input dataset column during Marketplace deployment?

[VincentF]

I have a marketplace product. This product has an input that is a Dataset with column “A, B,C”.
On installation of the marketplace product, I provide an input dataset that has only column “A,B”.

Marketplace is throwing an error, as column “C” is not found. However I know that column C is not important for this product. Hence I would like to override this default behavior and force the installation even if this column is not present in the input dataset I provide.

Is it possible to force marketplace to install on an input dataset that is missing a column ?

[mhogberg]

Hi @VincentF,

There is no way to force Marketplace to install the product without column C. You have two options today:

1. Add column C to your input dataset so that you get a complete input mapping
2. If you’re using Pipeline Builder, you can mark column C as optional in the source pipeline by right clicking on the graph → packaging settings → select input dataset → uncheck column C. You will then need to repackage your DevOps product.

[VincentF]

Thanks for this response !
Is there a way to tag a column as optional if generated from code repository ?

[mhogberg]

Unfortunately not, this is currently only supported in Pipeline Builder. What you can do as a workaround is put a pipeline that does an identity transform in front of your code repository transform, and mark the column as optional in the pipeline.