Best practices for creating csv file(s) with multiple record types

I am trying to build a csv file format for a complex type that follows a structure similar to a Car (A) which has Wheels (B) and Speakers (C), wherein each of these entities (Car, Wheel, Speaker) will also have their own properties specific to them (such as Car's color, Wheel's Air Pressure, Speaker's decibel).
I wanted to know if there are general best practices with regard to organizing multiple record types (in this case these 3 entities) in a csv format which make extracting data from the file simpler and error-free.
Should I create 1 file per record type or can group all of these 3 in single file?
If I put them in single file, should i organize them together per entity type (A-A-A-B-B-B-C-C-C) or per object (A-B-C-A-B-C-A-B-C)?

The csv file(s) will be loaded daily as a batch into a SQL database that has these 3 entities relationship in a normalized structure.