Ah ok, so i have to build a hash index to get all relevant data sets first, than cassa performs the range scan.

Good to know, thx a lot! :o)

2010/11/14 Jonathan Ellis <jbellis@gmail.com>
Because 0.7.0 indexes are more like a Hash index than a B-tree.

On Sun, Nov 14, 2010 at 12:39 PM, André Fiedler
<fiedler.andre@googlemail.com> wrote:
> Ok, i read this before. Could you explain (short) why i have to do this? In
> my opinion it isn't necessary, i want to understand why it is. ;o)
> thx André!
>
> 2010/11/14 Nate McCall <nate@riptano.com>
>>
>> You must have a clause with an EQ operator on an indexed column present.
>> See:
>> http://wiki.apache.org/cassandra/API07#IndexClause
>> and:
>> http://wiki.apache.org/cassandra/API07#get_indexed_slices
>>
>> For more details.
>>
>>
>>
>> On Sun, Nov 14, 2010 at 12:09 PM, André Fiedler
>> <fiedler.andre@googlemail.com> wrote:
>> > Hi, i wrote a question on the phpcassa group, but i think its more
>> > cassandra
>> > related. Would be nice, if you get some time and take a look:
>> >
>> > http://groups.google.com/group/phpcassa/browse_thread/thread/1b6acb5f7dccb94f
>> > greetings André
>
>



--
Jonathan Ellis
Project Chair, Apache Cassandra
co-founder of Riptano, the source for professional Cassandra support
http://riptano.com