You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@hbase.apache.org by "Edward J. Yoon" <ed...@apache.org> on 2009/02/03 10:01:10 UTC
Hbase client
I would like to have abstraction layers for the HTable and RowResult
because My project uses the two classes to implement Graph and AdjList
(or Matrix and Vector). w/o abstraction layers, the codes are becoming
very messy.
What do you think?
--
Best Regards, Edward J. Yoon @ NHN, corp.
edwardyoon@apache.org
http://blog.udanax.org
Re: Hbase client
Posted by "Edward J. Yoon" <ed...@apache.org>.
We may want to handle the RDF graph using Hbase, then If we use the
neighbors(String v) which returns a adjacency list of strings instead
of HTable.getRow(String row), we can move into more intuition-style
programming even if It wont related with best algorithms.
> but am not sure how best to port the idea to HBase yet.
Check this out -
http://blog.udanax.org/2009/01/distributed-matrix-multiplication-with.html
- you may receive some help.
On Wed, Feb 4, 2009 at 1:09 PM, Morton Swimmer <mo...@gmail.com> wrote:
>
> Besides the HBase instability, I am in any case still digesting the SHER
> approach to scaling RDF store and query. In principle, I like it, but am not
> sure how best to port the idea to HBase yet.
>
> Cheers, Morton
>
> On 3 Feb 2009, at 20:17, Edward J. Yoon wrote:
>
>>
>> I see. Thanks for your advice.
>>
>> On Wed, Feb 4, 2009 at 2:13 AM, Jim Kellerman (POWERSET)
>> <Ji...@microsoft.com> wrote:
>>>
>>> I would wait for HBASE-880 which will change the client API
>>> in a way that I think would be easier to create the abstractions
>>> you are looking at.
>>>
>>> ---
>>> Jim Kellerman, Powerset (Live Search, Microsoft Corporation)
>>>
>>>
>>>> -----Original Message-----
>>>> From: edward@udanax.org [mailto:edward@udanax.org] On Behalf Of Edward
>>>> J.
>>>> Yoon
>>>> Sent: Tuesday, February 03, 2009 1:01 AM
>>>> To: hbase-dev@hadoop.apache.org; hama-dev@incubator.apache.org
>>>> Subject: Hbase client
>>>>
>>>> I would like to have abstraction layers for the HTable and RowResult
>>>> because My project uses the two classes to implement Graph and AdjList
>>>> (or Matrix and Vector). w/o abstraction layers, the codes are becoming
>>>> very messy.
>>>>
>>>> What do you think?
>>>> --
>>>> Best Regards, Edward J. Yoon @ NHN, corp.
>>>> edwardyoon@apache.org
>>>> http://blog.udanax.org
>>>
>>>
>>
>>
>>
>> --
>> Best Regards, Edward J. Yoon @ NHN, corp.
>> edwardyoon@apache.org
>> http://blog.udanax.org
>
>
--
Best Regards, Edward J. Yoon @ NHN, corp.
edwardyoon@apache.org
http://blog.udanax.org
Re: Hbase client
Posted by "Edward J. Yoon" <ed...@apache.org>.
We may want to handle the RDF graph using Hbase, then If we use the
neighbors(String v) which returns a adjacency list of strings instead
of HTable.getRow(String row), we can move into more intuition-style
programming even if It wont related with best algorithms.
> but am not sure how best to port the idea to HBase yet.
Check this out -
http://blog.udanax.org/2009/01/distributed-matrix-multiplication-with.html
- you may receive some help.
On Wed, Feb 4, 2009 at 1:09 PM, Morton Swimmer <mo...@gmail.com> wrote:
>
> Besides the HBase instability, I am in any case still digesting the SHER
> approach to scaling RDF store and query. In principle, I like it, but am not
> sure how best to port the idea to HBase yet.
>
> Cheers, Morton
>
> On 3 Feb 2009, at 20:17, Edward J. Yoon wrote:
>
>>
>> I see. Thanks for your advice.
>>
>> On Wed, Feb 4, 2009 at 2:13 AM, Jim Kellerman (POWERSET)
>> <Ji...@microsoft.com> wrote:
>>>
>>> I would wait for HBASE-880 which will change the client API
>>> in a way that I think would be easier to create the abstractions
>>> you are looking at.
>>>
>>> ---
>>> Jim Kellerman, Powerset (Live Search, Microsoft Corporation)
>>>
>>>
>>>> -----Original Message-----
>>>> From: edward@udanax.org [mailto:edward@udanax.org] On Behalf Of Edward
>>>> J.
>>>> Yoon
>>>> Sent: Tuesday, February 03, 2009 1:01 AM
>>>> To: hbase-dev@hadoop.apache.org; hama-dev@incubator.apache.org
>>>> Subject: Hbase client
>>>>
>>>> I would like to have abstraction layers for the HTable and RowResult
>>>> because My project uses the two classes to implement Graph and AdjList
>>>> (or Matrix and Vector). w/o abstraction layers, the codes are becoming
>>>> very messy.
>>>>
>>>> What do you think?
>>>> --
>>>> Best Regards, Edward J. Yoon @ NHN, corp.
>>>> edwardyoon@apache.org
>>>> http://blog.udanax.org
>>>
>>>
>>
>>
>>
>> --
>> Best Regards, Edward J. Yoon @ NHN, corp.
>> edwardyoon@apache.org
>> http://blog.udanax.org
>
>
--
Best Regards, Edward J. Yoon @ NHN, corp.
edwardyoon@apache.org
http://blog.udanax.org
Re: Hbase client
Posted by Morton Swimmer <mo...@gmail.com>.
Besides the HBase instability, I am in any case still digesting the
SHER approach to scaling RDF store and query. In principle, I like it,
but am not sure how best to port the idea to HBase yet.
Cheers, Morton
On 3 Feb 2009, at 20:17, Edward J. Yoon wrote:
>
> I see. Thanks for your advice.
>
> On Wed, Feb 4, 2009 at 2:13 AM, Jim Kellerman (POWERSET)
> <Ji...@microsoft.com> wrote:
>> I would wait for HBASE-880 which will change the client API
>> in a way that I think would be easier to create the abstractions
>> you are looking at.
>>
>> ---
>> Jim Kellerman, Powerset (Live Search, Microsoft Corporation)
>>
>>
>>> -----Original Message-----
>>> From: edward@udanax.org [mailto:edward@udanax.org] On Behalf Of
>>> Edward J.
>>> Yoon
>>> Sent: Tuesday, February 03, 2009 1:01 AM
>>> To: hbase-dev@hadoop.apache.org; hama-dev@incubator.apache.org
>>> Subject: Hbase client
>>>
>>> I would like to have abstraction layers for the HTable and RowResult
>>> because My project uses the two classes to implement Graph and
>>> AdjList
>>> (or Matrix and Vector). w/o abstraction layers, the codes are
>>> becoming
>>> very messy.
>>>
>>> What do you think?
>>> --
>>> Best Regards, Edward J. Yoon @ NHN, corp.
>>> edwardyoon@apache.org
>>> http://blog.udanax.org
>>
>>
>
>
>
> --
> Best Regards, Edward J. Yoon @ NHN, corp.
> edwardyoon@apache.org
> http://blog.udanax.org
Re: Hbase client
Posted by Morton Swimmer <mo...@gmail.com>.
Besides the HBase instability, I am in any case still digesting the
SHER approach to scaling RDF store and query. In principle, I like it,
but am not sure how best to port the idea to HBase yet.
Cheers, Morton
On 3 Feb 2009, at 20:17, Edward J. Yoon wrote:
>
> I see. Thanks for your advice.
>
> On Wed, Feb 4, 2009 at 2:13 AM, Jim Kellerman (POWERSET)
> <Ji...@microsoft.com> wrote:
>> I would wait for HBASE-880 which will change the client API
>> in a way that I think would be easier to create the abstractions
>> you are looking at.
>>
>> ---
>> Jim Kellerman, Powerset (Live Search, Microsoft Corporation)
>>
>>
>>> -----Original Message-----
>>> From: edward@udanax.org [mailto:edward@udanax.org] On Behalf Of
>>> Edward J.
>>> Yoon
>>> Sent: Tuesday, February 03, 2009 1:01 AM
>>> To: hbase-dev@hadoop.apache.org; hama-dev@incubator.apache.org
>>> Subject: Hbase client
>>>
>>> I would like to have abstraction layers for the HTable and RowResult
>>> because My project uses the two classes to implement Graph and
>>> AdjList
>>> (or Matrix and Vector). w/o abstraction layers, the codes are
>>> becoming
>>> very messy.
>>>
>>> What do you think?
>>> --
>>> Best Regards, Edward J. Yoon @ NHN, corp.
>>> edwardyoon@apache.org
>>> http://blog.udanax.org
>>
>>
>
>
>
> --
> Best Regards, Edward J. Yoon @ NHN, corp.
> edwardyoon@apache.org
> http://blog.udanax.org
Re: Hbase client
Posted by "Edward J. Yoon" <ed...@apache.org>.
I see. Thanks for your advice.
On Wed, Feb 4, 2009 at 2:13 AM, Jim Kellerman (POWERSET)
<Ji...@microsoft.com> wrote:
> I would wait for HBASE-880 which will change the client API
> in a way that I think would be easier to create the abstractions
> you are looking at.
>
> ---
> Jim Kellerman, Powerset (Live Search, Microsoft Corporation)
>
>
>> -----Original Message-----
>> From: edward@udanax.org [mailto:edward@udanax.org] On Behalf Of Edward J.
>> Yoon
>> Sent: Tuesday, February 03, 2009 1:01 AM
>> To: hbase-dev@hadoop.apache.org; hama-dev@incubator.apache.org
>> Subject: Hbase client
>>
>> I would like to have abstraction layers for the HTable and RowResult
>> because My project uses the two classes to implement Graph and AdjList
>> (or Matrix and Vector). w/o abstraction layers, the codes are becoming
>> very messy.
>>
>> What do you think?
>> --
>> Best Regards, Edward J. Yoon @ NHN, corp.
>> edwardyoon@apache.org
>> http://blog.udanax.org
>
>
--
Best Regards, Edward J. Yoon @ NHN, corp.
edwardyoon@apache.org
http://blog.udanax.org
Re: Hbase client
Posted by "Edward J. Yoon" <ed...@apache.org>.
I see. Thanks for your advice.
On Wed, Feb 4, 2009 at 2:13 AM, Jim Kellerman (POWERSET)
<Ji...@microsoft.com> wrote:
> I would wait for HBASE-880 which will change the client API
> in a way that I think would be easier to create the abstractions
> you are looking at.
>
> ---
> Jim Kellerman, Powerset (Live Search, Microsoft Corporation)
>
>
>> -----Original Message-----
>> From: edward@udanax.org [mailto:edward@udanax.org] On Behalf Of Edward J.
>> Yoon
>> Sent: Tuesday, February 03, 2009 1:01 AM
>> To: hbase-dev@hadoop.apache.org; hama-dev@incubator.apache.org
>> Subject: Hbase client
>>
>> I would like to have abstraction layers for the HTable and RowResult
>> because My project uses the two classes to implement Graph and AdjList
>> (or Matrix and Vector). w/o abstraction layers, the codes are becoming
>> very messy.
>>
>> What do you think?
>> --
>> Best Regards, Edward J. Yoon @ NHN, corp.
>> edwardyoon@apache.org
>> http://blog.udanax.org
>
>
--
Best Regards, Edward J. Yoon @ NHN, corp.
edwardyoon@apache.org
http://blog.udanax.org
RE: Hbase client
Posted by "Jim Kellerman (POWERSET)" <Ji...@microsoft.com>.
I would wait for HBASE-880 which will change the client API
in a way that I think would be easier to create the abstractions
you are looking at.
---
Jim Kellerman, Powerset (Live Search, Microsoft Corporation)
> -----Original Message-----
> From: edward@udanax.org [mailto:edward@udanax.org] On Behalf Of Edward J.
> Yoon
> Sent: Tuesday, February 03, 2009 1:01 AM
> To: hbase-dev@hadoop.apache.org; hama-dev@incubator.apache.org
> Subject: Hbase client
>
> I would like to have abstraction layers for the HTable and RowResult
> because My project uses the two classes to implement Graph and AdjList
> (or Matrix and Vector). w/o abstraction layers, the codes are becoming
> very messy.
>
> What do you think?
> --
> Best Regards, Edward J. Yoon @ NHN, corp.
> edwardyoon@apache.org
> http://blog.udanax.org