Apache Kylin : Analytical Data Warehouse for Big Data

Page tree

Welcome to Kylin Wiki.


Page History

DateAuthorComment
2021-01-21For Kylin 4.0.0-beta .



Background

Before Kylin 4.0.0-beta, because of the fact that :

  • System Cube is no longer supported, so we cannot collect cuboid hit history;
  • Cuboid statistics is also not supported, so we cannot statistics row count of each cuboid;

so Kylin 4.0.0-alpha do not support Cube Planner.

But from Kylin 4.0.0-beta, Kylin team add Cube Planner Phase One support because we add Cuboid statistics in Cube Build Job.

Prepare

Step 1 : Enable cube planner

Since Kylin 4.0.0-beta only support Cube Planner Phase one, and it is not enable by default. You need to modify kylin.properites , and add kylin.cube.cubeplanner.enabled .

Step 2 : Modify cube level configuation

You can specific cube level properties when create a cube. This step is optional

Following is the (161) cuboid which left after cuboid purning by static rules.


Step 3 : Build you segment

After first segment be build, you will you cuboid list be updated to less because of Cube Planner phase one.

Attention

Attention, because of system cube did not provided cuboid level's query history, so cube planer phase two is not supported at the moment.

 

Reference



  • No labels