Re: Continuous queries with aggregations

[Ilya Kasnacheev <il...@gmail.com>]

Hello!

I am afraid you will have to implement dynamic aggregation yourself since
Continuouos Queries do not support any fashion of expressions, they just
return changed rows.

E.g. you can keep track of current min and max values for desired fields
and update these values if you have a new record (or no longer have an
existing one).

Regards,
-- 
Ilya Kasnacheev


чт, 27 сент. 2018 г. в 21:47, Francesco di Siena <ef...@gmail.com>:

> We have an ignite system with a table of about 100 columns, and 300,000
> rows.
> Of this data around 10 columns are changing frequently, i.e. up to several
> times per second on most of the 300K rows.We are able to query this data (e
> .g aggregation and grouping) at a reasonable speed (<1s).
>
> We can also use a "continuous query" to listen to all changes to the raw
> data, however this only notifies the subscriber of changes to individual
> table rows, not the aggregated results.
> Is there any mechanism to allow a continuous aggregated query?
>
> e.g. Say my query was:
>      select max(price), min(size), name from Products where price>23 group
> by name
> e.g. I'd like the client to receive updates at the aggregated level
> whenever
> max(price) changes due to an underlying row causing this to change.
>
> One work-around is to notice when a changed price was the current max, and
> if so, re-query the aggregated data, but this is not efficient.
> Is there any such a mechanism, or any suggested work-around?
> thanks
>
>
>
> --
> Sent from: http://apache-ignite-users.70518.x6.nabble.com/
>