Just realized my reply didn't include the riak user list. Forwarding it now.
Hi Luke, I'm assuming you're talking about Yokozuna (Riak Search 2.0). Answers > inline... > Sorry, I should have mentioned that I am using Riak 1.4.2 > Yes. You can POST data to the /yz/extract URL to see how your JSON (or > XML) is parsed by the extractor. > > > 2. If I want to return more than just key/value pairs what options do I > have aside from using search queries as input for MapReduce jobs? > > When search returns data it can return values as well, depending on > the Solr schema being used. Ideally you build your schema to index and > store just the data needed by your application. > In some instances the entire object is required, is there another option aside from storing the each key/value pair of the object? > > > 3. When it comes to "children" objects, is it better to search for > objects or rely on numerous links? > > I'm not sure what you mean here. Generally, child objects can be > stored either as part of their parent (nested) in a value, or as > values in another bucket. In the latter case, the parent object just > has a reference to the child's ID (and vice-versa, if necessary). The > decisions you make when modeling your data should take access, update > and search patterns into account, since all impact the shape of the > objects you store. > The "relationship" I'm attempting to model is a one-to-many, where the "children" have pieces of the "parent." I was previously considering "linking" each "child" to the "parent" and link walking when I need all the children. Alternatively, I was thinking of searching for the "parent_id" in the "children". Is one approach recommended over the other? Thanks. -Ari
_______________________________________________ riak-users mailing list riak-users@lists.basho.com http://lists.basho.com/mailman/listinfo/riak-users_lists.basho.com
