Using RData without zipping RRS feed

  • Question

  • I found out that the performance of my Azure Machine Learning Webservice largely depends on the size of the zip-file that I upload in which my prediction models are stored. I wonder whether there is a way to use an RData dataset, which can be uploaded through the datasets section, without needing to zip the file. My question is how this can be done, because my Execute R script does not allow to connect with an RData dataset? Or doesn't this improve the calculation time of the prediction?
    Wednesday, June 15, 2016 1:01 PM


All replies

  • Performance doesn't make any difference with the zip file. If you have packages that you installing in your zip file, they need to be installed by Azure ML.

    Thursday, June 16, 2016 7:30 PM
  • Though I still find it weird that I can upload RData files in Azure ML as a data set (without zipping) but cannot use them, I found out why the performance was bad. It was not the unzipping that took a lot of time, it was the loading of the RData in the Execute R script (load("*.RData")). I broke down the RData file into small pieces to just load the model that I needed and that improved the performance largely.
    Thursday, June 23, 2016 6:21 AM