Hi Everyone. Just checking if this use case can be satisfied in a sane way: data we don't currently have but know a lot about the metadata. We have two situations at the moment that result in this question. 1. Purchasing a dataset, knowing completely what we need it for, and how it will integrate with our current pipelines. 2. Data product design - understanding the edges of the data product, creating the data contracts etc. I was thinking of just making a fake file placeholder - e.g. add some header columns to a one line csv file. Wondering if there's a better way to think about these situations.

@bulky-soccer-26729 Would love your input here!