kylin-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From ShaoFeng Shi <shaofeng...@apache.org>
Subject Re: When cube modified,can we query from all new and history segment without purge data?
Date Thu, 01 Jun 2017 07:05:14 GMT
Check hybrid model: https://kylin.apache.org/blog/2015/09/25/hybrid-model/

2017-06-01 15:03 GMT+08:00 suheng.cloud <permanent_star@hotmail.com>:

> Hi, all
> In reality, some small modify may be add into existing cube(like add a new
> measure),I search for some infomations that we could clone a new cube, but
> in this way, we must also rebuild all history data in new cube, or we can't
> query the sql in which part_date includes old and new values.
>
> For example,kylin_sales_cube(partition start form 2012-01-01) build from
> 2012-01-01 to 2012-06-01
> and
> kylin_sales_cube_clone(partition start form 2012-09-01) build from
> 2012-09-01 to 2012-12-01
>
> from now on,we want to stop incremental build for the formmer but change to
> build the latter one.(since more and more cube will hard to manager and
> take
> more compute and storage resource?)
>
> sql:select part_dt, sum(price) as total_selled, count(distinct seller_id)
> as
> sellers from kylin_sales where part_dt > '2012-01-01'  group by part_dt
> order by part_dt;
>
> only show up statistic from kylin_sales_cube
>
> How can we make the sql show all data(merge result of 2012-01-01~2012-06-01
> and 2012-09-01~2012-12-01) without rebuild all history segment?
>
>
>
> Thanks a lot.
>
> --
> View this message in context: http://apache-kylin.74782.x6.
> nabble.com/When-cube-modified-can-we-query-from-all-new-and-
> history-segment-without-purge-data-tp8138.html
> Sent from the Apache Kylin mailing list archive at Nabble.com.
>



-- 
Best regards,

Shaofeng Shi 史少锋

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message