0

I build a time series model in microsoft analysis service 2008 r2. There is no problem to create the model and process. But when I make a pridiction query, there is NaN in the result. How can I avoid this? If there is no way to avoid the NaN, how can I save the result to SQL server? I used SQLBulkCopy class to insert the data into SQL table for better performance. Any suggestion is appreciate but need have good performance. The predication result is big. It has around 5 milion rows of data.

Does anybody have a good suggestion?

hoss
  • 2,430
  • 1
  • 27
  • 42
user2968062
  • 1
  • It might help if you provide more details about the query – Leeor Nov 08 '13 at 08:43
  • The query is very simple: SELECT FLATTENED [Actual_Base_1_1].[DIMKEY],PREDICTTIMESERIES([Actual_Base_1_1].[Actual],1,12) AS FORECAST FROM [Actual_Base_1_1]. The actual_base_1_1 is a time series model name. When you execute this query, there are a few rows contains NaN in the return result. The NaN will cause it fails when you try to save the result into a sql server table. – user2968062 Nov 13 '13 at 07:57

0 Answers0