Hi,

You want to use pre-aggregate datamaps, please try the example
`org.apache.carbondata.examples.PreAggregateDataMapExample` to understand
how it works. Please note that we are going to obsolete pre-aggregate
datamaps from next version and make MV(materialized view ) datamaps going
to replace it.

Regards,
Ravindra.

On Thu, 11 Jul 2019 at 17:40, Deepak Kulkarni <[email protected]> wrote:

> Thx. However we want to tame sure that the queries are using pre-aggregate
> datamaps . We tried using explain plan but the output showed in this link
> https://carbondata.apache.org/datamap-management.html we are not getting.
>
> Can you help?
>
> BR,
> Deepak
>
> On Tue, Jul 9, 2019 at 8:51 AM Ravindra Pesala <[email protected]>
> wrote:
>
>> Hello,
>>
>> 1. Currently there is no way to impose primary key indexes on carbon. We
>> may consider it in future.
>>
>> 2. We have datamaps interface opened for implementing secondary indexes.
>> Right now we have implemented for min/max , bloom indexes. User can add
>> there own implementations as well.
>>
>> 3. There is no constraints supported on carbon right now.
>>
>> 4. Spark/hive has many predefined functions and also user can have there
>> own user defined function and register to spark and use them carbon.
>>
>> Regards,
>> Ravindra.
>>
>> On Tue, 2 Jul 2019 at 2:52 PM, Deepak Kulkarni <[email protected]>
>> wrote:
>>
>>> Hi,
>>>
>>>
>>>
>>> We are evaluating the Carbondata for following alternative to Oracle
>>> –and hit following issues when achieving this migration -
>>>
>>>
>>>
>>>    1. Primary Key Indexes – how we can avoid duplicate records on
>>>    Carbon?
>>>    2. Secondary Indexes
>>>    3. Constraints such as NULL, NOT NULL – how can we set this at
>>>    column levels in carbon?
>>>    4. Functions such as DECODE – what alternatives we have in Carbon to
>>>    use these SQL functions?
>>>
>>>
>>>
>>> Appreciate the response.
>>>
>>>
>>>
>>> --
>>> Regards,
>>> Deepak
>>>
>> --
>> Thanks & Regards,
>> Ravi
>>
>
>
> --
> Regards,
> Deepak
>


-- 
Thanks & Regards,
Ravi

Reply via email to