Error when importing custom fields

[harryjr]

Hello,

I am sure this is not a formatting issue. I previously exported all of my custom fields into CSV. After deleting them, I have tried to import them back again. I have also tried to create a simple YAML file with one object to import, but get the same error:

<class 'AttributeError'>

'NoneType' object has no attribute 'app_label'

Python version: 3.9.13
NetBox version: 3.4.5

[harryjr]

Here is the simple .yaml file I am using to test:

VLAN_Domain_ID:
name: VLAN_Domain_ID
type: text
description: ID for VLAN Domain
required: true
search_weight: 0
filter_logic: Loose
ui_visibility: read-write
content_types: ipam.vlangroup

[sisayhaile]

I run into a similar issue when importing the custom field.
<class 'AttributeError'>

'NoneType' object has no attribute 'app_label'

Python version: 3.10.6
NetBox version: 3.5.3

[gurubert]

+1

[abhi1693]

I've not tested it but content types should be a list

[candlerb]

Here is the simple .yaml file I am using to test

YAML is indentation-sensitive. Please edit your post and put three backticks (```) on a line of their own, before and after the yaml, so that it can be copied accurately.

However, it seems to me that the main problem is the spurious outer tag (the first line VLAN_Domain_ID:). The following YAML works for me with Netbox v3.6.0; I also had to add "weight", and change "Loose" to "loose", but there were helpful error messages telling me when these were wrong.

name: VLAN_Domain_ID
type: text
description: ID for VLAN Domain
required: true
weight: 100
search_weight: 0
filter_logic: loose
ui_visibility: read-write
content_types: ipam.vlangroup

[kuhball]

Sadly just ran into this.

Python version: 3.11.4
NetBox version: 3.6.0
Plugins:
netbox_initializers: 3.6.0
netbox_prometheus_sd: 0.6

Import Data:

name;content_types;label;group_name;type;required;description;default;search_weight;fliter_logic;ui_visibility;Choice Set;
vlan_group_route_distinguisher;ipam.vlangroup;Route Distinguisher for EVPN / VXLAN;;Integer;FALSE;;;1000;Loose;Read/write;;

Stack trace from sentry:

AttributeError:'NoneType' object has no attribute 'app_label'
File "django/core/handlers/exception.py", line 55, in inner
     response = get_response(request)
File "django/core/handlers/base.py", line 197, in _get_response
     response = wrapped_callback(request, *callback_args, **callback_kwargs)
File "django/views/generic/base.py", line 104, in view
     return self.dispatch(request, *args, **kwargs)
File "netbox/views/generic/base.py", line 77, in dispatch
     return super().dispatch(request, *args, **kwargs)
File "utilities/views.py", line 99, in dispatch
     return super().dispatch(request, *args, **kwargs)
File "django/views/generic/base.py", line 143, in dispatch
     return handler(request, *args, **kwargs)
File "netbox/views/generic/bulk_views.py", line 456, in post
     new_objs = self.create_and_update_objects(form, request)
File "netbox/views/generic/bulk_views.py", line 414, in create_and_update_objects
     if model_form.is_valid():
File "django/forms/forms.py", line 201, in is_valid
     return self.is_bound and not self.errors
File "django/forms/forms.py", line 196, in errors
     self.full_clean()
File "django/forms/forms.py", line 433, in full_clean
     self._clean_fields()
File "django/forms/forms.py", line 445, in _clean_fields
     value = field.clean(value)
File "django/forms/models.py", line 1580, in clean
     value = self.prepare_value(value)
File "utilities/forms/fields/csv.py", line 118, in prepare_value
     return content_type_identifier(value)
File "utilities/utils.py", line 464, in content_type_identifier
     return f'{ct.app_label}.{ct.model}' 

[candlerb]

There are a lot of errors in the CSV you provided.

If Netbox generates an exception instead of a proper error message in the case of invalid data, then certainly that's a bug; it's not validating input properly. You can provide examples of such bad data and the corresponding exceptions, and stick them in a ticket.

However, the general issue here is that Netbox's CSV import format is (sadly) not the same as its CSV export format - this is documented in the FAQ. Therefore, you can't use CSV as a way to preserve data or move it between Netbox instances, and you can't use a CSV export as a template for a CSV import.

Here are the bugs I found in your CSV (which I also changed to comma-separated rather than semicolon):

1. You have bad column heading fliter_logic [sic], and also "Choice Set" instead of "choice_set". Netbox silently ignores such columns, which IMO is also a bug, but that is being worked on under CSV upload silently ignores invalid columns #13200
2. You are missing the mandatory weight (display weight) column. Mandatory columns are listed in bold in the CSV upload page. Note that both weight and search_weight are mandatory
3. You are using invalid values, e.g. "Integer" should be "integer", "FALSE" should be "false" (*), "Loose" should be "loose", "Read/write" should be "read-write". Again, an artefact of CSV import and export being arbitrarily different.
4. You have a spurious blank column at the end

Fixing those gets you to here:

name,content_types,label,group_name,type,required,description,default,weight,search_weight,filter_logic,ui_visibility,choice_set
vlan_group_route_distinguisher,ipam.vlangroup,Route Distinguisher for EVPN / VXLAN,,integer,false,,,100,1000,loose,read-write,

And now it works. (Note: if the content_type is "ipam.vlan_group" then you get "Record 1 content_types: this field is required", which is misleading since a value is provided, and if it's "ipam:vlangroup" then a different exception is raised: "not enough values to unpack (expected 2, got 1)")

Anyway I've given you a solution that works, so don't question it too much :-)

(*) Boolean columns do accept invalid values, and I think "FALSE" does actually work, but for example "fflse" is accepted as "true", and the documented allowed values are "true" and "false"

EDIT: made some corrections - I was using an invalid content type originally, and this confused me.