Problems submittimg azure machine learning pipelines

Question

I have a problem submitting my azure machine learning pipeline. I used the Import Data model to import data from a ULR via HTTP. When I clicked on submit, it submitted successfully with no green vertical line indication usually on the left side of the rectangular Import Data model. However, I got a green tick on the left pain that the job is completed. To continue building the pipeline, I added Select Columns in Dataset model to select the required columns. But as you can see from the screen shots, the select columns by name option is disabled and there are no names to select from.
When I click the job details on the left pain on the Authoring page, it opens a read-only pipeline which cannot be edited. When I clone it, I am able to edit it but I am unable to add Select Columns in Dataset model to select the required columns as previously.
I never experienced this problem with the Azure machine learning (Classic) or earlier versions of the Microsoft Azure Machine Learning Studio. I started experiencing this as soon as Microsoft changed the interface recently. Can someone please help me?

Answer 1

@Ebenezer Danso-Amoako Yes, with the change in UI the functionality to load the data from the downstream module is not being displayed, in this case it is select columns from dataset.
You can however enter the column names during authoring by first previewing the data from the import data module and note down the column names that needs to be added in downstream module.

We will pass the feedback of this change to the product team for review. Thanks!!

If an answer is helpful, please click on or upvote which might help other community members reading this thread.

Answer 2

please update your tutorials !

Answer 3

please update your tutorials !

Answer 4

I'm experiencing the same issue. After selecting the columns to clean, the column node does not seem to pull through the metadata of the column names. The two components I'm using are "Clean Missing Data" and "Select Columns in Dataset," and no matter which one I place first, the metadata is not passed through to the second one. This is certainly something that should work out of the box, and it has cost me quite a few hours by now. I am currently evaluating this product for use with a large client in a production environment, and when simple things like this do not work, the level of trust is not very high. Please reply and fix it ASAP.

Thank you,