Correct use of Editor-Only data

Development Platform & Builds
,
1

I’ve created a system where some of the objects in my game use custom scene components that have some additional UPROPERTY’s (some FVectors). Rather than fill these values out manually for every object, I wrote a script in blender that exports a CSV file with them alongside the corresponding FBX. The scene components have a UDataTable* (wrapped in WITH_EDITORONLY_DATA), and when PostEditChangeChainProperty detects a change to it, it populates the other non-editor-only UPROPERTY’s with the values from that file.

My question is, is this a good practice? And do I have to do anything with those CSV files to ensure they aren’t compiled into the production build?

2

I am doing similar thing.

I created blueprintable C++ function library. Then i create const arrays or variables there and functions that read that and expose to blueprints.

This way i am sure my default values are hardcoded is C++, and it is easy to modify array written in text, easier than copy / paste to blueprints or use scripts to change some values. And c++ does not get wiped out when you change struct names.