Re: non-indexable SOLR schema
Uhm, interesting, I didn't see that page; I believe the one I saw was about using Riak Search and that's where my assumption on needing to create an index to make a bucket_type came along. Much easier though! Thanks! Alex On Saturday, April 4, 2015, Shawn Debnath wrote: > Riak does not in any way rely on solr for its KV operations. Not sure > where you are seeing that or what code you are looking at but you can > define bucket types, activate them without ever touching solr. The basic > bucket type set up instructions can be found here: > http://docs.basho.com/riak/latest/dev/advanced/bucket-types/. > > On 4/4/15, 8:51 AM, "Alex De la rosa" > wrote: > > How to do that? > > For what I have seen, to use these special data types you need to create > a bucket under a bucket_type; and this bucket type is created following an > index so data gets indexed in Solr. > > On Saturday, April 4, 2015, Shawn Debnath > wrote: > >> You do not have to set up Yokozuna (yz) to be able to use counters, >> sets, maps in Riak. Those types can be used independently via the Riak key >> value store. You only need to set up the indexes if you were to search for >> data via yz or search 2.0. And in that case, you can remove the generic >> mappings, keep the _yz specific ones and then explicitly add the fields you >> want to index/search on. >> >> On 4/4/15, 7:25 AM, "Alex De la rosa" wrote: >> >> Hi all, >> >> To be able to use counters/sets/maps in Riak I have to store the object >> into a defined bucket_type indexed via SOLR. >> >> However, this will require extra disk space as data will be indexed (if >> using the default schema). Can I create a custom schema "ignoring" all >> fields so nothing is indexed? I don't need to use Riak Search on these >> objects, as I always know the KEY to fetch them. A schema like this would >> work? ( I guess that I can not put [ indexed="false" ] in the "_yz*" fields >> required by Riak, right? Or is it possible to not index that data either? ) >> >> —+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+— >> >> >> >> >> >> > multiValued="false" required="true"/> >>> multiValued="false"/> >>> multiValued="false"/> >>> multiValued="false"/> >>> multiValued="false"/> >>> multiValued="false"/> >>> multiValued="false"/> >>> multiValued="false"/> >>> multiValued="false"/> >> >> >> _yz_id >> >> >> > multiValued="true" class="solr.StrField" /> >> >> > sortMissingLast="true" /> >> >> >> —+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+— >> >> Another question, can I use the [ compressed="true" ] to save disk >> space? in both and "_yz*" fields? >> >> Thanks, >> Alex >> >> ___ riak-users mailing list riak-users@lists.basho.com http://lists.basho.com/mailman/listinfo/riak-users_lists.basho.com
Re: non-indexable SOLR schema
Riak does not in any way rely on solr for its KV operations. Not sure where you are seeing that or what code you are looking at but you can define bucket types, activate them without ever touching solr. The basic bucket type set up instructions can be found here: http://docs.basho.com/riak/latest/dev/advanced/bucket-types/. On 4/4/15, 8:51 AM, "Alex De la rosa" mailto:alex.rosa@gmail.com>> wrote: How to do that? For what I have seen, to use these special data types you need to create a bucket under a bucket_type; and this bucket type is created following an index so data gets indexed in Solr. On Saturday, April 4, 2015, Shawn Debnath mailto:sh...@debnath.net>> wrote: You do not have to set up Yokozuna (yz) to be able to use counters, sets, maps in Riak. Those types can be used independently via the Riak key value store. You only need to set up the indexes if you were to search for data via yz or search 2.0. And in that case, you can remove the generic mappings, keep the _yz specific ones and then explicitly add the fields you want to index/search on. On 4/4/15, 7:25 AM, "Alex De la rosa" > wrote: Hi all, To be able to use counters/sets/maps in Riak I have to store the object into a defined bucket_type indexed via SOLR. However, this will require extra disk space as data will be indexed (if using the default schema). Can I create a custom schema "ignoring" all fields so nothing is indexed? I don't need to use Riak Search on these objects, as I always know the KEY to fetch them. A schema like this would work? ( I guess that I can not put [ indexed="false" ] in the "_yz*" fields required by Riak, right? Or is it possible to not index that data either? ) —+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+— _yz_id —+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+— Another question, can I use the [ compressed="true" ] to save disk space? in both and "_yz*" fields? Thanks, Alex ___ riak-users mailing list riak-users@lists.basho.com http://lists.basho.com/mailman/listinfo/riak-users_lists.basho.com
Re: non-indexable SOLR schema
How to do that? For what I have seen, to use these special data types you need to create a bucket under a bucket_type; and this bucket type is created following an index so data gets indexed in Solr. On Saturday, April 4, 2015, Shawn Debnath wrote: > You do not have to set up Yokozuna (yz) to be able to use counters, > sets, maps in Riak. Those types can be used independently via the Riak key > value store. You only need to set up the indexes if you were to search for > data via yz or search 2.0. And in that case, you can remove the generic > mappings, keep the _yz specific ones and then explicitly add the fields you > want to index/search on. > > On 4/4/15, 7:25 AM, "Alex De la rosa" > wrote: > > Hi all, > > To be able to use counters/sets/maps in Riak I have to store the object > into a defined bucket_type indexed via SOLR. > > However, this will require extra disk space as data will be indexed (if > using the default schema). Can I create a custom schema "ignoring" all > fields so nothing is indexed? I don't need to use Riak Search on these > objects, as I always know the KEY to fetch them. A schema like this would > work? ( I guess that I can not put [ indexed="false" ] in the "_yz*" fields > required by Riak, right? Or is it possible to not index that data either? ) > > —+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+— > > > > > > multiValued="false" required="true"/> > multiValued="false"/> > multiValued="false"/> > multiValued="false"/> > multiValued="false"/> > multiValued="false"/> > multiValued="false"/> > multiValued="false"/> > multiValued="false"/> > > > _yz_id > > > multiValued="true" class="solr.StrField" /> > > /> > > > —+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+— > > Another question, can I use the [ compressed="true" ] to save disk > space? in both and "_yz*" fields? > > Thanks, > Alex > > ___ riak-users mailing list riak-users@lists.basho.com http://lists.basho.com/mailman/listinfo/riak-users_lists.basho.com
Re: non-indexable SOLR schema
You do not have to set up Yokozuna (yz) to be able to use counters, sets, maps in Riak. Those types can be used independently via the Riak key value store. You only need to set up the indexes if you were to search for data via yz or search 2.0. And in that case, you can remove the generic mappings, keep the _yz specific ones and then explicitly add the fields you want to index/search on. On 4/4/15, 7:25 AM, "Alex De la rosa" mailto:alex.rosa@gmail.com>> wrote: Hi all, To be able to use counters/sets/maps in Riak I have to store the object into a defined bucket_type indexed via SOLR. However, this will require extra disk space as data will be indexed (if using the default schema). Can I create a custom schema "ignoring" all fields so nothing is indexed? I don't need to use Riak Search on these objects, as I always know the KEY to fetch them. A schema like this would work? ( I guess that I can not put [ indexed="false" ] in the "_yz*" fields required by Riak, right? Or is it possible to not index that data either? ) —+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+— _yz_id —+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+— Another question, can I use the [ compressed="true" ] to save disk space? in both and "_yz*" fields? Thanks, Alex ___ riak-users mailing list riak-users@lists.basho.com http://lists.basho.com/mailman/listinfo/riak-users_lists.basho.com
non-indexable SOLR schema
Hi all, To be able to use counters/sets/maps in Riak I have to store the object into a defined bucket_type indexed via SOLR. However, this will require extra disk space as data will be indexed (if using the default schema). Can I create a custom schema "ignoring" all fields so nothing is indexed? I don't need to use Riak Search on these objects, as I always know the KEY to fetch them. A schema like this would work? ( I guess that I can not put [ indexed="false" ] in the "_yz*" fields required by Riak, right? Or is it possible to not index that data either? ) —+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+— _yz_id —+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+—+— Another question, can I use the [ compressed="true" ] to save disk space? in both and "_yz*" fields? Thanks, Alex ___ riak-users mailing list riak-users@lists.basho.com http://lists.basho.com/mailman/listinfo/riak-users_lists.basho.com
Re: [Announcement] Official Riak Node.js client released.
Thanks Brian for your reply and trying to work out the links part. I took a look at the RiakObject and could see how to manipulate it with FetchValue and the Secondary Index. However, I failed to figure out how to access the meta while reading an object. Just as an example, when you're updating an object you could be using some meta data, and some of them come from a prior read. If you could just share some examples with code that'd be very helpful. Thanks José On Apr 2, 2015, at 3:18 PM, Brian Roach wrote: > Jose, > > See: http://basho.github.io/riak-nodejs-client/classes/RiakObject.html > > As noted in the docs for StoreValue and FetchValue, you can use > RiakObject instead of a plain string or JS object, and metadata is the > specific reason. > > That said, it would appear I forgot about links as link walking is a > deprecated feature in Riak 2.0. I apologize for the oversight and will > add them. > > As for explicit examples beyond the normal API docs, our various pages > on docs.bacho.com are being updated to include node.js and will be > done soon. > > Thanks! > - Brian Roach > > > > > > On Thu, Apr 2, 2015 at 5:47 AM, Jose G. Quenum > wrote: >> >> Dear Brian, >> Thank you very much for sharing this link. I have been looking for a riak >> client for node.js that is compatible with riak 2.0. >> However, having glanced at it I noticed that there was no method to access >> the meta information. This is good for link manipulation, for example. As >> well in order to update an object the meta information could be useful. As a >> disclaimer here, I should mention that I have been using riak-js to access >> riak 1.4. Now I'd like to transition to riak 2.0 and I am looking for the >> right tools. >> >> Overall, where can one find more documentation and more concrete examples >> about how to manipulate riak with this client. >> Thanks & regards, >> Jose >> >> Sent from my iPad >> >>> On Apr 2, 2015, at 2:36 AM, Brian Roach wrote: >>> >>> Greetings Riak Users! >>> >>> Today we released the official Node.js client for Riak. >>> >>> It's available via npm: >>> >>> https://www.npmjs.com/package/basho-riak-client >>> >>> The github repo can be found here: >>> >>> https://github.com/basho/riak-nodejs-client >>> >>> API docs are published here: >>> >>> http://basho.github.io/riak-nodejs-client/classes/Client.html >>> >>> Thanks! >>> - Brian Roach >>> >>> ___ >>> riak-users mailing list >>> riak-users@lists.basho.com >>> http://lists.basho.com/mailman/listinfo/riak-users_lists.basho.com ___ riak-users mailing list riak-users@lists.basho.com http://lists.basho.com/mailman/listinfo/riak-users_lists.basho.com