September 19, 2016 at 2:44 pm
I've read up recently that some people use the Data Mining Query Task to output time series data into a relational table, I haven't had much luck hitting a series of NaN, Infinity etc errors, I am assuming div/0 errors. This is the output of the ARIMA / Blended Algorithm.
The only way I was able to do this is create a series of data flow task and within them run the MDX query and use a data conversion task into a destination container. I made all the data types in the relational table nvarchar max. Is there a better way to do this, how would you handle this type of task?
I would lke to use the data mining query container and have it dump into a source table, but when you set it up to create table the numerics default to float, when the NaN or something not numeric related hits it kills the ETL.
Are there containers that can transform NaN or whatever to 0? Hoping to get a clean solutions and learn something, thanks all.
Thanks for any input, I have been researching but nothing that addresses best practices in regards to this.
Thanks
Querying Microsoft SQL Server 2012/2014 - Certified
September 19, 2016 at 3:04 pm
Are there containers that can transform NaN or whatever to 0? Hoping to get a clean solutions and learn something, thanks all.
Assuming 'NaN' is a string, the way I would do this is to configure the source column in SSIS as a string, and then to use a Derived Column which performs the conversion (to INT/NUMERIC) and NaN clean-up for you.
It is then the Derived Column which you would map to your (properly typed) target column.
If you haven't even tried to resolve your issue, please don't expect the hard-working volunteers here to waste their time providing links to answers which you could easily have found yourself.
September 19, 2016 at 9:28 pm
Thanks for the follow up appreciate it. I'll give it a go tomorrow.
Querying Microsoft SQL Server 2012/2014 - Certified
Viewing 3 posts - 1 through 2 (of 2 total)
You must be logged in to reply to this topic. Login to reply