Samstag, 26. Januar 2013 07:41
Hi, I'm working on the samples provided with the book "Data Mining with SQL Server® 2008".
Sometimes to build a model there is only the DMX and not the associated BIDS project, so, to better understand I build the models using the DMX and I did a reverse engineering to a BIDS project.
It seems that in DMX you need only to define a Data Source, and not a Data Source View, so that this one is not found on on the BIDS project (I've got using reverse engineering). Since the models do not use a Data Source View I cannot in BIDS see the tables associated to them, and MORE IMPORTANT (!!!) I cannot process them.
I tried to hack the model so that it "knows" a DSV, but without success.
This is important in Time Series where I'm unable to reproduce in BIDS the features I can use in DMX. So it seems to me that the only way to build Time Series models is using DMX.
Thanks for the help.
Montag, 28. Januar 2013 02:26Moderator
We can specify a data source, but can not specify a data source view. Please refer to the following link about time series prediction DMX tutorial: http://technet.microsoft.com/en-us/library/38ea7c03-4754-4e71-896a-f68cc2c98ce2(v=sql.100)
If you have any feedback on our support, please click here
TechNet Community Support
Mittwoch, 30. Januar 2013 14:48
I think what teodorom wants is to run the model in BIDS. I have a same issue which is still unsolved. I imported a database created in DMX into BIDS and was not able to process the mining structure due to an "structure does not contain bindings to data" error.
So our issue is to know how to bind data to our mining structure (and create the associate data source view) and connect our existing models to it.
I guess a similar situation would apply to our issue - how to bind an existing mining structure to a new data source and data source view assuming the columns setting are the same? (For example, we have data from two different sources and want to add the second one for the test after the structure and model are already built for the first one.)