# Bundle Data Types

### Bundle Data Presentation Types

| Kind                                                                     | Type                           | Example                                                                                                                                                                      |
| ------------------------------------------------------------------------ | ------------------------------ | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| literal or **file link**                                                 | int, float, str, bool, unicode | <p><code>return 3</code><br><code>return 3.14</code><br><code>return "pi"</code><br><code>return True</code></p><p><code>return "s3://abucket/akey/afile.parquet"</code></p> |
| None type                                                                | NoneType                       | `return None`                                                                                                                                                                |
| <p>flat collection of</p><p>literals <strong>and file links</strong></p> | list, tuple                    | <p><code>return \[1,3,5]</code><br><code>return ("jd", "ke")</code></p>                                                                                                      |
| <p>dictionary of flat</p><p>collections</p>                              | dictionary                     | `return {'name':["mark","sue"], 'age':[21, 32]}`                                                                                                                             |
| **file links**                                                           | str or `luigi.LocalTarget`     | <p><code>return /Users/kyocum/fancy.csv</code></p><p><code>return s3://abucket/akey/afile.parquet</code></p>                                                                 |
| dataframe                                                                | `pandas.DataFrame`             | return pd.DataFrame({'name':\['mark','sue']})                                                                                                                                |

{% hint style="danger" %}
Currently Disdat only supports simple dataframes (no hierarchical indices or complex object types).  &#x20;

If you want to store something complex, there are many good file formats that already exist: csv, parquet, orc, etc.    Choose one of those and store the link!
{% endhint %}