On 22/07/2011 17:56, Patrick Julien wrote:
> I can still use it for keys if I don't need ranges then? Because for
> what we are doing we can always re-assemble keys
yes,but why would you use CompositeType if you don't need range query?
> On Fri, Jul 22, 2011 at 11:38 AM, Donal Zang<zangds@ihep.ac.cn> wrote:
>> If you are using OPP, then you can use CompositeType on both key and column
>> name; otherwise(Random Partition), just use it for columns.
>> On 22/07/2011 17:10, Patrick Julien wrote:
>>> With the current implementation of CompositeType in Cassandra 0.8.1,
>>> is it recommended practice to try to use a CompositeType as the key?
>>> Or are both, column and key, equally well supported?
>>>
>>> The documentation on CompositeType is light, well non-existent really,
>>> with
>>>
>>> key_validation_class set to CompositeType (UUIDType, IntegerType)
>>>
>>> can we query all matching rows just by using CompositeType(UUIDType)?
>>>
>>> In my specific use case, what would work best is to have a composite
>>> key that is a CompositeType with thousands of columns each.
>>>
>>
>> --
>> Donal Zang
>> Computing Center, IHEP
>> 19B YuquanLu, Shijingshan District,Beijing, 100049
>> zangds@ihep.ac.cn
>> 86 010 8823 6018
>>
>>
>>
--
Donal Zang
Computing Center, IHEP
19B YuquanLu, Shijingshan District,Beijing, 100049
zangds@ihep.ac.cn
86 010 8823 6018
|