Kylin can do this (by defining "price" as dimension) but the cube explosion rate might be high since "price"'s cardinality is high;
Usually when face such a scenario we can optimize it in this way: 1. define a hive view, adding another column say "price_grade"; normalizing the price values into grade, say 100, 200, etc (depends on the need); 2. using the new view as cube's fact table, adding "price_grade" as a dimension, and defining "sum(price)" as measure; 3. query using the new added column as the filtering condition. 2016-12-21 15:30 GMT+08:00 Alberto Ramón <[email protected]>: > I never use, but Kylin 976 > <https://issues.apache.org/jira/browse/KYLIN-976> can be useful for you > > > 2016-12-21 8:14 GMT+01:00 ZhouJie <[email protected]>: > >> hi, everyone >> i want to know if kylin can filter a column which has been measured, as >> follows: >> select sum(price) from hotprice_copy1 where price > 100.0 and price >> <5000.0 >> >> thanks >> joe >> > > -- Best regards, Shaofeng Shi 史少锋
