wal | Developer - Big Data(I&D)| 9731648376 | www.mu-sigma.com
>
> From: Pierre Chalamet [mailto:pie...@chalamet.net]
> Sent: Monday, May 28, 2012 3:31 PM
> To: user@cassandra.apache.org
> Subject: Re: Doubt in Row key range scan
>
> Hi,
>
> It's normal.
&g
oubt in Row key range scan
Hi,
It's normal.
Keys to replicas are determined with a hash (md5) when using the random
partitionner (which you are using I guess).
You probably want to switch to the order preserving partionner or tweak your
data model in order to rely on 2nd index for such
You are using the Random Partitioner.
Using the RP is a good thing because you avoid hot spots, but it has
its defaults too. You can't scan a slice of row, they won't be ordered
because all your keys are stored using their md5 values.
You should review your data model to use columns to order your
Hi,
It's normal.
Keys to replicas are determined with a hash (md5) when using the random
partitionner (which you are using I guess).
You probably want to switch to the order preserving partionner or tweak your
data model in order to rely on 2nd index for such filtering.
- Pierre
-Origin