1. Join fact table and lookup tables to be big flat intermediate table in hive. 2. Extract distinct values on columns that require dictionary. 3. Build dictionary for columns. 4. Build base cuboid from intermediate table, apply dictionary encoding. 5. Take iterations to aggregate every cuboid. Starting from the base cuboid, aggregate off one dimension in each iteration. 6. Cut HBase regions by scanning the key space of all cuboids. 7. Create HBase table. 8. Load all cuboids into HBase table.
On Tue, Mar 17, 2015 at 5:13 PM, koteswara Vemu <[email protected]> wrote: > Hi, > > I am trying to understand the work flow of Cube Build Action in Kylin. > Just wanted to go through the flow and understanding what each step of > build action is performing to my data. But unfortunately end up with lack > of documentation. Can you please explain the steps to understand each step > of cube build action? > > Thanks and Regards, > Koti >
