So between a start and an end date, how is Kylin going to keep a track of new data and perform incremental build? Is the job for building an incremental cube triggered periodically or manually by the user?
On Mon, Apr 6, 2015 at 3:38 PM, Shi, Shaofeng <[email protected]> wrote: > If you¹re going to incrementally build a cube, you need specify the > ³partition column² when create the cube, and give it a start date; Then > when you run it, Kylin will ask you to enter an end date; Kylin will use > this start date and end date as condition to generate the intermediate > flat table. You can run a Kylin sample to get it. > > Regarding your questions: the date column need be on fact table, while it > is not a must that the date be a dimension. > > On 4/6/15, 1:46 PM, "Abhishek Sinha" <[email protected]> wrote: > > >Hi all, > > > >How does incremental cube build work in Kylin? Does the date dimension > >need > >to be present in fact and each of the dimension? > >What changes need to be made at the warehouse/star schema level in order > >to > >support incremental build? > > > > > >Thanks > >-- > >Abhishek Sinha > >Mobile: +919035191078 > >infoworks.io > > -- Abhishek Sinha Mobile: +919035191078 infoworks.io
