A hive table may have 100 columns (what you see in ³Table² tap), if you build 10 columns into a cube (dimension or measure), you will just see the 10 columns in the ³Query² tab for that table;
That¹s to say, the columns in ³Query² tap are the available columns for query with Kylin; Based on your previous description, if you drop the cube test1, as there are other READY cubes supporting this table, it will still in the query tab. On 5/12/15, 11:47 AM, "dong wang" <[email protected]> wrote: >thanks shaofeng, yang, yes, it puzzled me a lot, since I the structure >of >table test1 under "QUERY" table differs from that of the table test1 under >"Tables", but it is the same with structure of the CUBE "test1" I built, >thus, I think it as a CUBE till now~ > >in addition, if I drop the cube test1, what will shown under "QUERY" tab >for the syncd hive table "test1", what's the rule of the tables shown >under >"QUERY" tab? > >2015-05-12 11:39 GMT+08:00 Li Yang <[email protected]>: > >> I see the question now.. well, what lists on "Query" tab is tables not >> cubes. You have 10 cubes, but they all from one table. So for analyst >>(who >> don't care about cube), all they know is there's a table called test1, >> he/she can query. >> >> And yes, you can drop the test1 cube which does not contain any filter. >> >> On Tue, May 12, 2015 at 11:30 AM, dong wang <[email protected]> >> wrote: >> >> > 0, I have a fact table named test1 >> > 1, I build a cube with no filter and the same name 'test1' as that of >>the >> > fact table, the cube 'test1' is shown on the "Query" tab >> > 2, then I build 10 little cube with FILTERs(different where) and with >> > different names based on the SAME fact tables test1, the names of 10 >> little >> > cubes seperately are : test1_sub1, test1_sub2, test1_sub3, ... , >> > test1_sub10, after all of the 10 cubes are built successfully, it >> shocked >> > me a lot, I cannot see any test1_subx under "Query" tab >> > >> > however, when I do some queries, I can find the log shows that it hit >>the >> > cube correctly which I want with the info "Realization Names:", which >> seems >> > satisfy what Luke talked to me last time, but I still wonder, can I >>drop >> > the above cube "test1", since I only want the query to hit test1_subx >> > >>
