I have been attempting to run the RevoScaleR_SqlServer_GettingStarted.R script that is part of the MS SQL Server 2016 package for a project I am working on. For some context, here is a PDF of the code that is provided by MS:
packages.revolutionanalytics.com/doc/8.0.0/win/RevoScaleR_SqlServer_Getting_Started.pdf
The code works well except for a strange error I receive when I attempt to run the rxHistogram() function as part of the tutorial. At step 3.2 of the tutorial it has me refine the data source by creating factors and identifying the min and max of specific variables. I run the rxGetVarInfo() function and receive the proper output as shown below:
The next step asks me to run the rxHistogram() function on the data and it seems to work well because it created the histogram in the pdf.
However, there is a consistent issue after I run this function that affects the running of the rest of the code, especially the rxCube() function which immediately follows. It seems that the rxHistogram() function removes the factors from my data. After some frustration, I re-ran the rxGetDataInfo() function and found my data had been changed.
NO FACTORS!!!
I would love to know if anyone else has seen or dealt with this issue and how it was solved? I have found the same issue with other rx() functions such as rxCube() so I feel there may be an issue in my environment setup.
