This browser is no longer supported.
Upgrade to Microsoft Edge to take advantage of the latest features, security updates, and technical support.
![](data:image/svg+xml, %3Csvg xmlns='http://www.w3.org/2000/svg' height='64' class='font-weight-bold' style='font: 600 30.11764705882353px "SegoeUI", Arial' width='64'%3E%3Ccircle fill='hsl(297.6, 46%, 38%)' cx='32' cy='32' r='32' /%3E%3Ctext x='50%25' y='55%25' dominant-baseline='middle' text-anchor='middle' fill='%23FFF' %3EKL%3C/text%3E%3C/svg%3E)
I have a model which I would like to test against a test dataset. I am using AutoML and have generated a regression model. I notice that when I register a data asset it randomly incorrectly assigns certain columns as Integer or Decimal. When I try to test the model against a new data asset there is a mismatch of about 50 columns. I have about 1400 columns in my dataset so it is impossible to go through every column to make sure they match what was in the orginal data asset used for the model. I don't even know how to view or change the schema of a data asset as it doesn't seem to allow me to do that in the Azure ML Studio. Is there a workaround for this problem or at least a way I can programatically update my test data asset schema to match the schema used in the model?
Thanks
@Kayode Leonard You can view and transform your data asset through Azure ML studio designer. Click on the Designer symbol from the left hand menu and use the Classic Prebuilt pipelines to create a new pipeline using classic prebuilt components. This should open up the canvas to view your data asset, drag and drop your data asset and use any of the data transformation components to transform your data.
For example, you can first view your data on the canvas using the Preview data option when you right click on your data component. You can then add the data transformation modules to remove missing columns or normalize data or any of the other transformation modules. Once the required logic is set in this module you can use the convert to dataset module to update or register a new data asset. This component also allows to set missing values and register a new data asset.
I hope this helps!!
Thanks!I realized the GUI was a pain and there are better options via pipelines, or using the python SDK.
Thanks!I realized the GUI was a pain and there are better options via pipelines, or using the python SDK.