> For the complete documentation index, see [llms.txt](https://docs.gaiodataos.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.gaiodataos.com/tools/tasks/analytics/sample.md).

# Sample

<figure><img src="/files/JqREMiCr23EB9kM1t0w3" alt=""><figcaption></figcaption></figure>

The **Sample** task in Gaio DataOS allows you to extract a subset of data from a table in a simple and controlled way. This functionality is ideal for **testing, validation, initial visualizations, or preprocessing in Machine Learning workflows**.

{% hint style="warning" %}
This task can only be used when a **table is selected** in the flow.️&#x20;
{% endhint %}

***

## How to Use the Sample Task

***

### 1. **Add the Sample Task to Your Flow**

* In the **Studio**, go to the **Tasks** panel.
* Under the **Analytics** section, select **Sample** task.

***

### 2. **Configure the Main Fields**

* **Task label**: (optional) Provide a name for this task within your flow.(default: `sample`)
* **Result table**: name of the output table that will contain the sampled data (e.g., `sample_sample`)

***

### 3. **Choose the Sampling Type**

You can choose between two options:

&#x20;**Percentage**

* Allows you to define the percentage of rows to be sampled from the original table.
* You can adjust the **slider** or manually input the value.
* **Example**: `0.7` (70%) → returns 70% of the rows from the source table.

**Rows**

* Allows you to define a **fixed number of rows** to extract as a sample.
* **Example**: `1,000` → the output table will contain exactly 1,000 randomly selected rows.

***

### 4. **Save and Execute**

* Once you’ve configured the sample type and value, click **Save**.
* Run the flow — a new table will be generated based on the selected sample configuration.

***

#### Best Practices

* Use the **Sample** task to:
  * Reduce dataset size during development or dashboard previews.
  * Create smaller datasets for training ML models.
  * Test queries and transformations without processing the full dataset.
* Combine with other tasks like **AutoML**, **Cluster**, or **Scoring** to streamline your experimentation and modeling.

{% hint style="success" %}
All columns from the source table will be present in the random table. Only the number of lines will be smaller.
{% endhint %}

***