Hi Tyler,
You're proposed design would work well with Riak. Buckets in Riak, by
default, are "free" meaning 100 buckets wouldn't consume any more resources
than 1 bucket.
Also, by using the date as the key you could map/reduce over a user's data
for set periods of time. For instance, consider the following entries:
/riak/user1/20100929
/riak/user1/20100930
/riak/user1/20101001
/riak/user1/20101002
/riak/user1/20101003
/riak/user1/20101004
To map/reduce over the first 3 days of October you can specify your
map/reduce inputs as:
{"inputs":[["user1","20101001"],["user1","20101002"],["user1","20101003"]],
...query goes here... }
Daniel Reverri
Developer Advocate
Basho Technologies, Inc.
[email protected]
On Wed, Sep 29, 2010 at 9:30 AM, Tyler Weir <[email protected]> wrote:
> Hello,
>
> I'm working on a side project and I don't know much about reasonable
> design for a Riak schema.
>
> I'll have a hybrid model, with user profiles sitting in Postgres and
> their data in Riak.
>
> If I were to store a blob of info per user per day (3 text blobs, a
> date, a URL and tags), does it make sense to have one bucket per user?
> If so, the key for the daily blob is simple, the date (unless that is
> also silly).
>
> I'm guessing 100 users, with maybe 400 RiakObjects per user per year.
>
> Does that design seem sound, or am I breaking some fundamental rules?
>
> Sorry for naivety,
> Tyler
>
>
> -------------------------------------
> http://www.tylerweir.ca
>
> _______________________________________________
> riak-users mailing list
> [email protected]
> http://lists.basho.com/mailman/listinfo/riak-users_lists.basho.com
>
_______________________________________________
riak-users mailing list
[email protected]
http://lists.basho.com/mailman/listinfo/riak-users_lists.basho.com
