Hi, just wonder if this is feasible. I have been using Qlikview and very like the way they slice cube data. You can save the fact data by snapshot date and save them into individual files. But I do not like each time I want to read data from files, I have to reload dimensions and related fact files into memory, which is not flexible, like SSAS.I understand SSAS also write sliced data into files. But as long as number of partitions grows very quickly and large, it is hard to manage. So I am thinking if there is a way to detach the old partition files, save them into other drives, and if customer would like to access the old data, we can attach related partition files into SSAS database again. I noticed <cube name>.<version>.det.xml file saves the partition infomation but any changes I perform (like delete one partition name) does not affect the SSAS database on the server (I tried to restart the SSAS server but it still won't work). I am guessing the only way to implement the change might be reprocess the cube. But this is so expensive.I also considered merging old partitions to the archived one. But the merging cost is so expensive. Any inputs or new ideas are very much appreciated!
↧