julia> at1 = Attributes(;c = 1)
> Attributes with 1 entry:
  c => 1

julia> at2 = Attributes(;c = 2)
> Attributes with 1 entry:
  c => 2

julia> merge(at1, at2)
> Attributes with 1 entry:
  c => 1

Where in the standard library as it states in the docs:

Construct a merged collection from the given collections. If necessary, the types of the resulting collection will be promoted to
accommodate the types of the merged collections. If the same key is present in another collection, the value for that key will be the
value it has in the last collection listed. See also mergewith for custom handling of values with the same key.

And it works as stated.

julia> d1 = Dict(:a => 1)
> Dict{Symbol, Int64} with 1 entry:
  :a => 1

julia> d2 = Dict(:a => 2)
> Dict{Symbol, Int64} with 1 entry:
  :a => 2

julia> merge(d1, d2)
> Dict{Symbol, Int64} with 1 entry:
  :a => 2

Wouldn't it be easier just to use the merge on the underlining Dict and keep it consistent with the expectation of people that are familiar with the merge function? If it is used somewhere deeper in the pipeline, it should not be an expansion of the merge method and be called something else... (You can assign me)