fieldPresets doesn't work properly

[socheatleang]

I am working with insert mutation which require the fieldPresets.

• I have ProfileLinks which have (id, link_id, profile_id, created_at, etc...)
• profile_id has exist inside the jwt-token sessionVariable: x-hasura-profile-id

So far, I would love to have fieldPresets with profile_id from the session variable.

---
kind: TypePermissions
version: v1
definition:
  typeName: InsertProfileMediasObject
  permissions:
    - role: user
      output:
        allowedFields:
          - createdAt
          - id
          - image
          - profileId
          - sequence
          - type
          - updatedAt
      input:
        fieldPresets:
          - field: profileId
            value:
              sessionVariable: "x-hasura-profile-id"

neither literal value inside the fieldPresets. I am sure that the x-hasura-profile-id has value.

Suprisely, which that data were insert into db is null instead of the id of profile.

Anyone meet this issue before, I am using hasura ddn v3.

[robertjdominguez]

@socheatleang — thanks for starting this discussion! I've run into a similar issue — though with ArgumentPresets — and solved it this way:

First, I created a native query:

INSERT INTO public."registrations" ("user_id", "ip_address", "lat", "long") VALUES (
  {{user_id}},
  {{ip_address}},
  {{lat}},
  {{long}}
)
RETURNING *

Then, generated the metadata for it and — as part of the CommandPermissions — added the ArgumentPreset using the session variable:

---
kind: CommandPermissions
version: v1
definition:
  commandName: insert_registration
  permissions:
    - role: admin
      allowExecution: true
    - role: user
      allowExecution: true
      argumentPresets:
        - argument: user_id
          value:
            sessionVariable: x-hasura-user-id

This then allows for a mutation such as:

mutation INSERT_REGISTRATION($ip_address: text!, $lat: text!, $long: text!) {
    insert_registration(ip_address: $ip_address, lat: $lat, long: $long) {
      affected_rows
    }
  }

IIRC, there's some issue with passing non-array values to an autogenerated mutation. @danieljharvey — is there anything you'd add or change from what I shared above?

[robertjdominguez]

Closing this due to inactivity.