Hey Benedict; I screwed up on email after a bachelor party, and sent something to external cassandra-users not internal users (drunken drivel)
I never said anything about it because I hoped no one noticed it.
That said, I was wondering if my data was helpful for your injector post. We
haven’t played with it yet, but I passed it on to some other Austin companies
who probably have a need.
Currently on hold with Orbitz - wtf - I opted to be called back because the
wait time was 40 mins; they called me back and I’ve still been on hold for 20
mins. I have a bunch of friends in LA - computer games/movies … my flight back
was routed by LA, so I figure I could leave Thursday night instead of Friday
morning and catch up with them all
> On May 23, 2015, at 2:13 AM, Benedict Elliott Smith
> <belliottsm...@datastax.com> wrote:
>
> Hi Min,
>
> The key selection occurs prior to this. The operation has been assigned one
> (or more, in the case of user profile operations) partition keys, and this
> is just it accessing that key. You should explore backwards for assignment
> operations, and see where these happen, to understand how this behaves.
> On 23 May 2015 01:30, "Min Zhou" <mz...@apache.org> wrote:
>
>> Hi all,
>>
>>
>>
>> Seems there is only one implementation of the getKey() , it's
>> in PredefinedOperation.java cassandra branch 2.2.0-beta1
>>
>>
>> protected ByteBuffer getKey()
>> {
>> return (ByteBuffer) partitions.get(0).getPartitionKey(0);
>> }
>>
>> The read operations will just the same key for each iteration, since it
>> will lead 100% cache hit on the storage side, the result throughput will be
>> very high.
>>
>> please correct me if i was wrong.
>>
>>
>> Min
>>
smime.p7s
Description: S/MIME cryptographic signature
