Search code examples
azuredatasetazure-machine-learning-service

how can I use data assets in azure ml registries in azure ml designer?

I have created some uri_folder data asset in my custom Azure ml registry. Also, I have created some components in my registry.

When I want to create a new custom pipeline in Azure ML Designer, I can use my custom component, but I can't see my registered data assets.

How can I use shared/Registered data assets in designer in multiple workspaces? Thanks.

Solution

  • Follow below steps to access dataset in multiple workspaces.

    Here, I created the data asset named sample in workspace jgsML as shown below.

    enter image description here

    Then created new workspace named jgsnew and created datastore linked to the previous workspace jgsML. Below is the storage account in jgsML.

    enter image description here

    With this storage account I created data store in workspace jgsnew.

    enter image description here

    Click on create > Give a name > select storage account under workspace having data asset.

    enter image description here

    Then select the blob container where the file is. To find this go data asset , there you will find storage uri as below, you can see under which container the file is there. select that blob container while creating datastore.

    enter image description here

    Next select Authentication type as Account key and add your storage account key.

    Below is the sample datastore I created.

    enter image description here

    After creating come to designer pipeline, drag and drop the Import Data component and Select Data source as datastore > select created datastore > select the path as below and click on preview schema to see columns.

    enter image description here

    And you will get the preview schema as below.

    enter image description here

    Like this you can access the data asset from multiple workspace.