> For the complete documentation index, see [llms.txt](https://docs.waveline.ai/extract/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.waveline.ai/extract/types/datashapeelement.md).

# DataShapeElement

Represents a single variable to be extracted from a document.

```ts
type DataShapeElement = {
  name: string;
  type: 'string' | 'number' | 'boolean';
  description?: string;
  isArray?: boolean;
} | {
  name: string;
  type: 'object';
  description?: string;
  isArray?: boolean;
  elements: DataShapeElement[];
};
```

* **`name`** (string): The name of the variable you want to extract.
* **`type`** ('string' | 'number' | 'boolean' | 'object'): The data type of the variable.
* **`description`** (string, optional): A brief description of the variable.
* **`isArray`** (boolean, optional): Indicates if the variable should be treated as an array.
* **`elements`** (DataShapeElement\[], required if `type` is 'object'): An array of elements describing the shape of the object.

### Notes on the `description`

We want to note that the **description** is optional but can greatly improve the result. It is used to give our AI a hint as to what you mean by the field that you want to extract.

#### Example

Let's say you name your field `first_name` within an E-Mail conversation. Without any further explanation, it might not be clear whether the recipient's or the sender's first name should be returned.