Passing multiple datasets in columnRole for a custom plugin

lina

Hi!

I am creating a plugin and I'm struglling on how to pass multiple datasets into the columnRole parametar in the JSON file. (I have two input datasets and I wanna have the option to select a column from either of the two and not just from one dataset):

  "inputRoles" : [
        {
            "name": "first_dataset",
            "label": "First Dataset",
            "arity": "UNARY",
            "required": true,
            "acceptsDataset": true
        },
        {
            "name": "second_dataset",
            "label": "Second Dataset",
            "arity": "UNARY",
            "required": true,
            "acceptsDataset": true
        }
    ]

and here is the column:

    {
      "name": "partitioning",
      "label": "unique identifier for a record",
      "description": "",
      "type": "COLUMN",
      "allowedColumnTypes": [
        "tinyint",
        "smallint",
        "int",
        "bigint",
        "string"
      ],
      "columnRole": "first_dataset",
      "mandatory": true
    },

Thanks!!

StanG

Hi,
There is no way to pass multiple datasets in the columnRole field of the same parameter.
Instead, you could for instance create a SELECT parameter to choose from which dataset to select a column from, use 2 different parameters with 2 different columnRole corresponding to the 2 input datasets and display the right parameter in the UI based on the SELECT parameter using a visibilityCondition.

This would look like this:

        {
            "name": "selected_dataset",
            "label": "Input dataset to select a column from",
            "type": "SELECT",
            "defaultValue": "first_dataset",
            "selectChoices": [
                {
                    "value": "first_dataset",
                    "label": "First dataset"
                },
                {
                    "value": "second_dataset",
                    "label": "Second dataset"
                }
            ]
        },
        {
            "name": "partitioning_first",
            "label": "unique identifier for a record",
            "description": "",
            "type": "COLUMN",
            "columnRole": "first_dataset",
            "visibilityCondition": "model.selected_dataset == 'first_dataset'"
        },
        {
            "name": "partitioning_second",
            "label": "unique identifier for a record",
            "description": "",
            "type": "COLUMN",
            "columnRole": "second_dataset",
            "visibilityCondition": "model.selected_dataset == 'second_dataset'"
        }

lina

@StanG
Thank you so much!