riak-users Digest, Vol 3, Issue 12

Aneil Mallavarapu aneil.mallavar at gmail.com
Thu Oct 15 10:53:28 EDT 2009


Riak looks truly awesome.  I'd like to stay tuned into the progress of
the indexing project.  Does it have a name yet?

Thanks for the reply,

Aneil

On Thu, Oct 15, 2009 at 7:36 AM, Justin Sheehy <justin at basho.com> wrote:
> On Thu, Oct 15, 2009 at 8:53 AM, Aneil Mallavarapu
> <aneil.mallavar at gmail.com> wrote:
>
>> 1. Is there a way to create ordered indexes in Riak, which are ~
>> btree-efficient for insert/update/modify of the indexed field?
>
> There is not yet any predefined functionality for this in Riak.  There
> is an ongoing project to add a few different kinds of indexed querying
> to Riak, but there is not yet a known schedule for public availability
> of that code.
>
>> 2. Is there a way to cluster a set of objects to the same nodes?  For
>> example, by providing a custom hash-function which returns the same value
>> for each element of the set.  Or does this not end up being an important
>> optimization to have to worry about?
>
> For general-purpose use, the default location hashing will provide
> what you want.  If you're really sure that you know why you ought to
> change it, you can override the default by setting the "chash_keyfun"
> bucket parameter.  If you change that parameter, store some data, then
> change it again, you might have confusing results.
>
> -Justin
>



More information about the riak-users mailing list