> ## Documentation Index
> Fetch the complete documentation index at: https://docs.membrane.agency/llms.txt
> Use this file to discover all available pages before exploring further.

# Data Schemas

Data Schema is used whenever you need to describe the data structure without providing the actual value.

Data Schema we use is based on [JSONSchema](https://json-schema.org/) with additional features.

## Additional Properties

### Reference Records (Enums or Picklists):

To specify reference records that can be used for a field (for example, to create a field with limited options for selection) use schema like this:

```yaml theme={null}
status:
  type: number # Type of `id` keys below
  referenceRecords:
    - name: Open # Title of that option displayed in the UI
      id: 1 # Actual value that will be used after selecting an option
    - name: Won
      id: 2
    - name: Lost
      id: 3
```

### Reference Data Table

You can point to a specific Data Table to use its records as possible values for a field.

Example:

```yaml theme={null}
ownerId:
  type: string
  referenceDataTable:
    dataTableId: dt_users
    label: name
    value: id
```

### Is Sensitive

There is an additional property `isSensitive` (boolean) that can be used to specify that the field contains sensitive information and should be treated accordingly (hidden in the UI, excluded from logs, etc).

Example:

```yaml theme={null}
password:
  type: string
  isSensitive: true
```

## Dynamic Data Schema

You can use [Formulas](/docs/references/formulas) in data schemas to make them dynamic. For example:

```yaml theme={null}
parent:
  type: object
  properties:
    type:
      type: string
      enum: ['tasks', 'projects']
    id:
      type: string
      referenceDataTable:
        dataTableId:
          $var: parent.type
```

In this, `parent.id` property will reference the task table if `parent.type` is `tasks` and the project table if `parent.type` is `projects`.

Schema will be evaluated in the context of the current value the schema is applied to.

It works in a special way for `array` schemas: when evaluating dynamic schemas, all arrays leading to the current value will be replaced with the current item that is being evaluated.

For example:

```yaml theme={null}
parents:
  type: array
  items:
    type: object
    properties:
      type:
        type: string
        enum: ['tasks', 'projects']
      id:
        type: string
        referenceDataTable:
          dataTableId:
            $var: parents.type
```

Notice the locator in the `id` field: `parents.type`. It treats `parents` as if it was an object rather than array. When evaluated, it will use the current item of the `parents` array.

For example, if the schema is applied to the following object:

```yaml theme={null}
parents:
  - type: tasks
    id: 1
  - type: projects
```

The schema for the first item of the `parents` list will be the following:

```yaml theme={null}
type: object
properties:
  type:
    type: string
    enum: ['tasks', 'projects']
  id:
    type: string
    referenceDataTable:
      dataTableId: tasks
```

But for the second item, it will be:

```yaml theme={null}
type: object
properties:
  type:
    type: string
    enum: ['tasks', 'projects']
  id:
    type: string
    referenceCollection:
      key: projects
```