locked
Error: config file is missing: '~/.azureml/settings.json' Traceback: RRS feed

  • Question

  • I am doing following steps in Azure Machine Learning studio which gives error in executing code in R Jupyter notebook

    These below steps

    Create a Jupyter Notebook
    1.  Browse to https://studio.azureml.net and sign in using the Microsoft account associated with
    free Azure ML account.
    2.  Create a new blank experiment, and name it Stats.
    3.  in the experiment items pane on the left, expand Saved datasets, expand Samples, and drag the
    Automobile price data (Raw) dataset onto the experiment canvas.
    4.  Right-click the output of the dataset, and in the menu, point to Open in a new Notebook, and
    then click R as shown here:

    when pressing run below

    library("AzureML")
    ws <- workspace()
    dat <- download.datasets(ws, "Automobile price data (Raw)")

    Below error are on my screen.

    Error: config file is missing: '~/.azureml/settings.json'
    Traceback:
    
    1. workspace()
    2. validate.AzureML.config(config, stopOnError = TRUE)
    3. stop(msg, call. = FALSE)
    
    

    where is the problem

    Sunday, September 23, 2018 8:31 AM

Answers

  • The problem is that the option "Open in a new Notebook" doesn't insert by itself the configuration data regarding to your experiment (for instance the workspace_id in workspace() ), that is needed for the Notebook to connect and work properly with AML Studio data.

    The best option is to use 'Generate Data Access Code...', available also from the menu at the output of the module. This option gives you the R code that you must copy and paste in your Notebook to access correctly to datasets or intermediate data in your experiment. You can paste this code in Notebooks and also in your R IDE.

    If the 'Generate Data Access Code...' option is disabled (it looks gray), drag a 'Convert to CSV' module and connect the output of the module from which you need the data to the input of this one. Run the experiment and then click on the 'Generate Data Access Code...' option depicted above, at the output of the 'Convert to CSV' module.

    Maybe these figures do help:


    • Proposed as answer by Ariel Cedola Sunday, September 23, 2018 9:32 AM
    • Marked as answer by Atifqur Sunday, September 23, 2018 3:19 PM
    Sunday, September 23, 2018 9:31 AM

All replies

  • The problem is that the option "Open in a new Notebook" doesn't insert by itself the configuration data regarding to your experiment (for instance the workspace_id in workspace() ), that is needed for the Notebook to connect and work properly with AML Studio data.

    The best option is to use 'Generate Data Access Code...', available also from the menu at the output of the module. This option gives you the R code that you must copy and paste in your Notebook to access correctly to datasets or intermediate data in your experiment. You can paste this code in Notebooks and also in your R IDE.

    If the 'Generate Data Access Code...' option is disabled (it looks gray), drag a 'Convert to CSV' module and connect the output of the module from which you need the data to the input of this one. Run the experiment and then click on the 'Generate Data Access Code...' option depicted above, at the output of the 'Convert to CSV' module.

    Maybe these figures do help:


    • Proposed as answer by Ariel Cedola Sunday, September 23, 2018 9:32 AM
    • Marked as answer by Atifqur Sunday, September 23, 2018 3:19 PM
    Sunday, September 23, 2018 9:31 AM
  • Thank you very much dear 
    Sunday, September 23, 2018 3:20 PM