simple subfields

[fostermh]

it seems to me there is support for repeating subfields but not subfields that are not repeating. I find it very convenient to organize the schema to support subfields.

example:

  {
      "field_name": "vertical-extent",
      "label": "Vertical Extent",
      "simple_subfields": [{
          "field_name": "min",
          "label": "Min",
          "form_attrs":{"pattern":"[0-9.]+"}
        },
        {
          "field_name": "max",
          "label": "Max",
          "form_attrs":{"pattern":"[0-9.]+"}
        }
      ]
    }

While this could be done with two fields it is more obvious to users that the min and max values are related when grouped as subfields.

[Mohab25]

@fostermh hi Matthew, is the simple_subfields added? when i use it, my inputs get disappeared

[fostermh]

Hi @Mohab25, I'm not sure what you mean? The pull request to merge this feature into master is still open so no it has not been added to the default branch if that is what you are asking. or do you mean you are testing #297 and it is not working for you?

[fostermh]

turns out the PR was broken. I think it is working now.

[ortegi]

Any update on using simple subfields? :(

[prohde]

Still no support for simple subfields?

[wardi]

simple subfields e.g.

{"my_group": {"my_field": …}}

clash with an assumption baked in to ckan's validation code that schemas will be validating alternating layers of dicts and lists e.g.

{"my_subfields": [
  {"my_field": …},
  …
]}

The way we define schemas we provide a validator for the latter (dict-list-dict) form as:

schema = {
  "my_subfields": {
    "my_field": [validators_go_here]
  }
}

There isn't a schema syntax to validate the former (dict-dict) form.

We would need to add a new schema syntax to ckan for simple subfields, possibly:

imaginary_future_schema = {
  ("my_group", "my_field"): [validators_go_here]
}

but this is outside of what's possible from ckanext-scheming at this point

[prohde]

I see. But wouldn't it be possible to set a maximum for the repeating fields causing the 'add new' to disappear?