How to create an oracle hierarchical query that transforms sparse data to dense data

Say you have a tree of data

A
|
+----B
|    |
|    +----C
|
+----D

to represent this in an oracle database we might have

create table mydata (id number(10), parent_id number(10), name varchar2(10))

and use a hierarchical query to traverse the tree.

What is the best way to define a sparse property set, such that child nodes inherit property values from the closest parent node?

In the above example, say we want to add owner and colour to the tree - how can the hierarchical query be modified to provide the inherited values?

My first thought was to create a new table

create table myproperties (id number(10), owner varchar2(30), colour varchar2(30));

We'd want the values in this table to be sparse as well - so a row in myproperties might specify 'owner' but leave 'colour' null. A given node in mydata would have a merged set of all non-null parent properties.

Then do an analytic query to fill in the gaps:

select * from (
    select
        id, groupid, parent_id, 
        last_value(owner ignore nulls) over (partition by groupid order by l desc) owner,
        last_value(colour ignore nulls) over (partition by groupid order by l desc) colour
    from
        (
            -- query from leaf to root
            select
                id, connect_by_root id groupid, level l, p.owner, p.colour
            from
                mydata d left outer join myproperties p on p.id=d.id
            start with
                id in (select id from mydata)
            connect by prior
                parent_id=id
        )
)
where
    id=groupid -- filter out redundant rows required by the analytic function

Is this a reasonable way to achieve the goal of a dense view on sparse hierarchical properties, or is there a better way (more efficient, more comprehensible)?