It's all there in the source. ;) Responses inline
[1]: https://github.com/DistributedNonsense/CorrugatedIron/blob/master/CorrugatedIron/Models/RiakObjectId.cs [2]: https://github.com/DistributedNonsense/CorrugatedIron/blob/master/CorrugatedIron.Tests.Live/RiakClientTests.cs#L62 --- Jeremiah Peschka - Founder, Brent Ozar Unlimited MCITP: SQL Server 2008, MVP Cloudera Certified Developer for Apache Hadoop On Sun, Feb 10, 2013 at 2:28 PM, Kevin Burton <[email protected]>wrote: > So for [2] what is a RiakObjectID? The name of the argument is bucket-key > pairs? But I am confused what is the constructor that takes an array of > strings that is labeled object id? No relation to RiakBucketKeyInput. > RiakObjectId is nothing more than a wrapper around a Bucket/Key pair [1] that are wired up for easy JSON conversion for Get operations. RiakBucketKeyInput is a sub-class of RiakPhaseInput for MR operations. > **** > > ** ** > > I hadn’t even known about IRiakBatchClient. Now I am intrigued. From the > source it seems to mimic all of the calls to IRiakClient. How is the setup > different? When does the “batch” start? > There are examples of IRiakBatchClient in the test suite [2] > **** > > ** ** > > This is kind of off-topic but I notice that there is a property public int > RetryCount { get; set; }. What is the default value? How does it work? Is > the whole batch retried or individual calls within the batch? > RetryCount is related to the cluster config that you set up in app.config - it's the number of retries before CorrugatedIron marks a node as dead. > **** > > ** ** > > *From:* riak-users [mailto:[email protected]] *On Behalf > Of *Jeremiah Peschka > *Sent:* Sunday, February 10, 2013 4:05 PM > *To:* riak-users > *Subject:* Re: CorrugatedIron batch get?**** > > ** ** > > My guess is that you're referring to the method Put(IEnumerable<RiakObject> > values) [1]. If that's the case, yes there's also a Get that accepts an > IEnumerable of RiakObjectID [2]. Riak doesn't support multi-get so CI is > just wrapping a bunch of Gets in the background for you. MapReduce might be > more effective here.**** > > ** ** > > If you're talking about the IRiakBatchClient interface [3], all actions of > RiakClient can be handled through IRiakBatchClient. Keep in mind that CI's > batch operations are just a wrapper around making a large number of calls > at once. Nothing about Batch() is atomic - if you need a Unit of Work > pattern you'll have to implement it yourself.**** > > ** ** > > [1]: > https://github.com/DistributedNonsense/CorrugatedIron/blob/master/CorrugatedIron/RiakClient.cs#L306 > **** > > [2]: > https://github.com/DistributedNonsense/CorrugatedIron/blob/master/CorrugatedIron/RiakClient.cs#L191 > **** > > [3]: > https://github.com/DistributedNonsense/CorrugatedIron/blob/master/CorrugatedIron/IRiakBatchClient.cs > **** > > > **** > > ---**** > > Jeremiah Peschka - Founder, Brent Ozar Unlimited**** > > MCITP: SQL Server 2008, MVP**** > > Cloudera Certified Developer for Apache Hadoop**** > > ** ** > > On Sun, Feb 10, 2013 at 1:32 PM, Kevin Burton <[email protected]> > wrote:**** > > Looking at the method calls I can see that there is a batch Put. Where I > can form a list of RiakObject s and do one put that writes all of the > objects to the database. Is there an equivalent “batch get”? Examples? > Thank you.**** > > > _______________________________________________ > 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
