I guess you might potentially have saved on key len building a secondary nested sets index or something along those lines as in http://en.wikipedia.org/wiki/Nested_set_model but simplicity of key path approach is hard to beat of course.
On Mon, Feb 21, 2011 at 10:46 AM, Vishal Kapoor <[email protected]>wrote: > Friends, > > I was deciding on table designs and came up with a problem where I needed > parent child relationship, a tree kind of dataset. > I did look a few presentations and slides and could not think in the same > way, it was more of HBase vs relational there... > > but here is what I did > > Grandpa : John > Grandma : Pam > Father : Peter > Mother : Julie > Aunt : Sandra > Sister : Rosy > Cousin : Rob > > Table > Family Tree > RowId comment > > John/Peter John's family > Pam/Peter Pam's family > John/Peter/Me John's family > Pam/Peter/Me Pam's family > Pam/Sandra/Rob Pam's family > Peter/Me Peter's family start here > Julie/Me Julie's family > Julie/Rosy Julie's family > ... > ... > > > the design works for me, and I am able to represent > Countries/States/Zips etc > > just wanted to get feedback from friends here... > > thanks, > Vishal Kapoor >
