You are viewing a plain text version of this content. The canonical link for it is here.

Posted to user@hbase.apache.org by Tux Racer <tu...@gmail.com> on 2009/12/10 09:49:29 UTC

client API newbie question: reverse order scanning

Hello Hbase Users,

I am reading the client API documentation and googling the Hbase users list:
it seems there is no way to create a scanner that will scan keys in the 
reverse order.
Has the implementation of reverse order scanning be considered already? 
Would that be something nice to have or just too complicated to 
implement or that would slow down Hbase?

Thanks in advance
TR

Re: client API newbie question: reverse order scanning

Posted by Tux Racer <tu...@gmail.com>.

Thanks Andrew for your feedback. I was already considering a second 
table to store the reversed order; so it is good to know this is the 
typical way to handle this :)
Thanks again
TR

Andrew Purtell wrote:
> As you have probably seen through your searches, the way this is typically
> handled is to create and maintain a secondary table with the keys reversed,
> such that scans on that table retrieve items in reverse order. This can be
> done in the application or using THBase. See 
>     http://hadoop.apache.org/hbase/docs/current/api/org/apache/hadoop/hbase/client/tableindexed/package-summary.html
> If using THBase, implement an IndexKeyGenerator with the desired properties
> and make sure it is on the classpath of all clients and region servers. 
>
>     - Andy
>
>
>
> ________________________________
> From: Tux Racer <tu...@gmail.com>
> To: hbase-user@hadoop.apache.org
> Sent: Thu, December 10, 2009 12:49:29 AM
> Subject: client API newbie question: reverse order scanning
>
> Hello Hbase Users,
>
> I am reading the client API documentation and googling the Hbase users list:
> it seems there is no way to create a scanner that will scan keys in the reverse order.
> Has the implementation of reverse order scanning be considered already? Would that be something nice to have or just too complicated to implement or that would slow down Hbase?
>
> Thanks in advance
> TR
>
>
>
>       
>

Re: client API newbie question: reverse order scanning

Posted by Andrew Purtell <ap...@apache.org>.

As you have probably seen through your searches, the way this is typically
handled is to create and maintain a secondary table with the keys reversed,
such that scans on that table retrieve items in reverse order. This can be
done in the application or using THBase. See 
    http://hadoop.apache.org/hbase/docs/current/api/org/apache/hadoop/hbase/client/tableindexed/package-summary.html
If using THBase, implement an IndexKeyGenerator with the desired properties
and make sure it is on the classpath of all clients and region servers. 

    - Andy



________________________________
From: Tux Racer <tu...@gmail.com>
To: hbase-user@hadoop.apache.org
Sent: Thu, December 10, 2009 12:49:29 AM
Subject: client API newbie question: reverse order scanning

Hello Hbase Users,

I am reading the client API documentation and googling the Hbase users list:
it seems there is no way to create a scanner that will scan keys in the reverse order.
Has the implementation of reverse order scanning be considered already? Would that be something nice to have or just too complicated to implement or that would slow down Hbase?

Thanks in advance
TR