Hi Billy , Thanks for your reply. I also found it in the jira https://issues.apache.org/jira/browse/KYLIN-1875 ,and I noticed that" This limitation is a problem for many-to-many relationships", but this is exactly where we are.
I tried to create a model with one fact table and another fact table as lookup table , there was an error in step 4:build dimension dictionary , and the information is "the lookup table dup key found". Now in model , I set the measure column to be the dimension , not any more measure , but in cube , I set the measure column to be the dimension and measure to query the raw data and measure of the measure column , because we want not just the intersection of the measure columns in two tables but the "count(distinct)" of them. Is this solution permissible ? Is there any problems or pitfalls ? Or is there a better solution ? Thanks. Best regards, Yang Gao -----邮件原件----- 发件人: Billy Liu [mailto:[email protected]] 发送时间: 2018年10月18日 18:36 收件人: user <[email protected]> 主题: Re: how to define a snowflake model In snowflake model, there is only one fact table also. But you could define lookup table joins with another lookup table. If the query pattern between your two "fact" tables is fixed, just like factA left join with factB. You could define factB as lookup table and skip the snapshot for this huge lookup table. With Warm regards Billy Liu 高扬02 <[email protected]> 于2018年10月18日周四 上午11:36写道： > > Kylin 2.5. > > How to define a snowflake model ? I did not find the place to set it. > > I have two fact tables , now I must create two cubes , and query “cube join > cube”. > > To query the raw data of the measure column for “join” sql ,Now I set the > measure column to be the dimension column also,because it takes too long to > use the “query-push-down”. > > Please show me where to define a snowflake model. > > thanks.
