MSSQL, OLAP, MDX, DMX: how to organize?

Question

OLAP

What is the point: I have a data plate in my Microsoft database. After spending quite a long time with the mouse in SQL Server Business Intelligence Development Studio (tm), I built an OLAP cube, on which I can now do all sorts of statistical things (build classification trees, regression trees, etc.). There is no problem to pull out, say, prediction on another (test) sample. The problem starts when, say, I want to save the results of the predictions, because access to OLAP is read-only, and a new cube with new data cannot be built there.
But the answer is: I just need to pull the OPENQUERY () function from normal Transact-SQL, and I can save the query data. Fine.
Now imagine that I applied a bunch of regressions or trees, got a new table (note that it is not in Analysis Services, but simply in the database). And now I need to apply trees again on the new table! So, I need to build a new OLAP cube, and all this should be done automatically, from a script.
Question: is it really possible to do this? Because the CREATE GLOBAL CUBE statement is really something wrong, it requires another cube as an argument.
UPD : as far as I understand, I do not need an OLAP cube, but a data source view. But how to create it from requests is also unclear)

Answer 1

CREATE GLOBAL CUBE - this is the creation of a local cube,
you can do this if you use XML / A, but it is not convenient and difficult to
have a third-party free script project - many use it www.bimlscript.com
recently talked about the Russian-language experience of using habrahabr.ru/post/ 259081 The presentation was titled "BIML is a SSIS Developer's Best Friend"