Just to add to the already excellent suggestions that have been given, enable 
PDX and only select the specific fields that you need and thus bypass 
deserialization. If you’re running inside a function, you may not even need the 
query but there’s not enough context to make any recommendation for or against.

Wes


On Jan 12, 2018, at 12:31 PM, Anthony Baker 
<[email protected]<mailto:[email protected]>> wrote:

You might want to explore CQRS approaches to optimize data storage based on 
access patterns.

Anthony


On Jan 12, 2018, at 8:49 AM, Amit Pandey 
<[email protected]<mailto:[email protected]>> wrote:

Actually I plan to do that, but do get the data I need to query.

My plan is to only get local data from Query (and I want to optimize this as 
much as possible).

Regards

On Fri, Jan 12, 2018 at 10:03 PM, Jacob Barrett 
<[email protected]<mailto:[email protected]>> wrote:
Amit,

You’re better off processing the data in place. Can you write your processing 
as a function that can execute on each node and process only the local data and 
returning only the results of that processing to your client?

If your going to continue with queries then yes, faster serialization can be 
achieved through implementing DataSetializable.

-Jake


> On Jan 11, 2018, at 11:28 PM, Amit Pandey 
> <[email protected]<mailto:[email protected]>> wrote:
>
> Hi Guys,
>
> I have a query which fetches some 30,000 rows in a single machine.
> I want to get under a second if possible or much lesser and do some 
> processing on it.
> I have some indexes is it possible to give me some tips on Optimizing 
> performance?
>
> Does implementing DataSerializable in domain class help ? I am using 
> ReflectionbasedSerializer.
>
> Regards



Reply via email to