Hi all,
We are wondering whether there are fundamental differences between Foundry Views and regular datasets in Foundry in terms of capabilities, i.e. can a View be used everywhere a dataset can be used (e.g. as input to a transform, to export to external systems etc.)?
In our use case, we would like to rely on the concept of Foundry Views to avoid data duplication and unnecessary identity transform computation between different projects. We are trying to assess whether we could accidentally corner ourselves into a problematic situation later down the line with this.
Thanks for your help!
A View dataset can be used in any downstream application like a regular dataset.
A View is a Foundry dataset that does not hold any files containing data, but returns the union of the backing datasets when it is read. View datasets build like regular foundry builds, but complete almost instantly since no data is actually read or written as part of the build.
There are a few possible use cases for using Views:
- Folder organization - Replicate a dataset in a different location, but without wasting compute and storage.
- Updating the input datasets to a union dataset without the need to modify code.
- Creating virtual datasets, such that the backing datasets can be changed in the background without disrupting existing analyses or pipelines.
1 Like