I’m facing some challenges in setting up my userdefined dataprep block. Mostly because of being inexperienced.
I want to design a block which performs multiple merge operations at once.
The currently provided dataprep blocks allow a single merge operation only. As I intend to have over 50+ merge operations per model, it will flood my dataprep reciepes.
The intention is to filter on a single metadata key, e.g. ue_merge, and perform the merge operation on different key values, e.g. item1, item2.
The metadata key is to be user provided. Unique metadata key-values are to be listed from the model.
The resulting merged items will have the key value as name.
My first question is, whether this should be a fetcher block or operation block?
Hello @Evmatica
No actual progress currently, but I think the solution will be comparable to the Distributing Static Mesh Actors example on the Create Custom Dataprep Blocks documentation page.
Also I’m still considering what will be most efficient way in terms of instancing.
I spoke to the guys from Epic who maintain Datasmith and Dataprep and both per actor for loops and support of the mesh merge function in the Dataprep context are not supported yet.
What they advised as a solution is to import everything as is, then use a custom Editor Utility Blueprint to loop over actors in the level and merge in the base Unreal Engine context vs in Dataprep.
Which is not so great in terms of having to import dirty, merge, then clean up old meshes but this looks like the only way for now.
Hi! I was wondering if there have been any changes to Dataprep capabilities in the time since this thread. I was hoping I could use Dataprep to individually merge contents of a bunch of objects based on metadata and hit an issue where the output is always a single mesh. If I’m reading the thread correctly, I believe you had the same issue?
Actually still no progress on this topic. I think in general there are no updates to dataprep?
Another missing aspects is spreading assets over multiple levels. As far as I know, one can only use one level per dataprep recipe.
Maybe ‘layers’ would resolve as a work around.