Hi Kylin Developers,

I'm confused about when to use derived dimension and how it will affects query 
performance. So far I've known that derived dimension will not be included in 
row key and value, then how kylin achieves to implement query grouping by these 
dimensions? The "Design Cube in Kylin" document says "Derived dimension reduces 
combination from 2N to 2 at the cost of extra runtime aggregation." I find it 
hard to understand.

Also, it's not clear to me what is the requirement for using derived dimension. 
In "Add Dimension" step, there is a tip saying "Using Derived for One-One 
relationship between columns, like ID and Name". Does it mean derived dimension 
should be unique key of the lookup table? If that's the case, the applicable 
scenario seams limited.

I'm aware that there's a JIRA https://issues.apache.org/jira/browse/KYLIN-887 
<https://issues.apache.org/jira/browse/KYLIN-887> for it, but I can't wait to 
know the answer :-)

Regards,
Dayue

Reply via email to