> ## Documentation Index
> Fetch the complete documentation index at: https://help.datafetcher.com/llms.txt
> Use this file to discover all available pages before exploring further.
# Data transformation
When you run a request, Data Fetcher converts the API response into a table of fields and records that can be written to your destination Airtable table. You can control how this data is structured using **response field mapping** and **data transformation** settings.
## Understanding the two transformation types
Data Fetcher offers two ways to handle nested arrays in your API responses: **Single record per item** and **Expand nested items**. The key difference is whether nested arrays stay **combined in one record** or are **expanded into multiple records**.
Let's use this sample JSON response to illustrate both approaches:
```json theme={null}
[
{
"id": "company1",
"positions": [{ "id": "position1" }, { "id": "position2" }]
},
{
"id": "company2",
"positions": [{ "id": "position3" }, { "id": "position4" }]
}
]
```
### Single record per item
**Single record per item** keeps each parent entity in one row by combining nested values with commas. Each company stays in its own record:
| **Id** | **Positions id** |
| -------- | ------------------- |
| company1 | position1,position2 |
| company2 | position3,position4 |
**Single record per item** doesn't mean the entire response becomes one record.
Data Fetcher finds the **first array in the response** and creates **one record for each item in that array**.
### Expand nested items
**Expand nested items** creates a new record for each item in a nested array. Parent fields are duplicated so each nested item has its own row.
| **Id** | **Positions id** |
| -------- | ---------------- |
| company1 | position1 |
| company1 | position2 |
| company2 | position3 |
| company2 | position4 |
## How to set the transformation mode
1. On the create screen, open **Advanced settings**.
2. Under **Data Transformation**, select either **Single record per item** or **Expand nested items**.
## Advanced Single record per item options
### Combine nested object keys
By default, nested object values are combined into a **comma-separated field**. Using our example above, all `positions.id` values become a single-line text field called **Positions id**.
You can disable this to create separate numbered fields instead:
| **Id** | **Positions 1 id** | **Positions 2 id** |
| -------- | ------------------ | ------------------ |
| company1 | position1 | position2 |
| company2 | position3 | position4 |
**To disable combining nested object keys:**
1. In advanced settings, ensure transformation is set to **Single record per item.**
2. Turn off **Combine nested object keys**.
### Combine arrays of text/numbers
Similarly, arrays of text or numbers are combined into a **comma-separated value** by default. For this response:
```json theme={null}
[
{
"id": "company1",
"categories": ["Fintech", "Startup"]
},
{
"id": "company2",
"categories": ["Healthtech", "Public", "Unicorn"]
}
]
```
The default output combines categories:
| **Id** | **Categories** |
| -------- | ------------------------- |
| company1 | Fintech,Startup |
| company2 | Healthtech,Public,Unicorn |
You can disable this to create separate numbered fields:
**To disable combining arrays:**
1. In advanced settings, ensure transformation is set to **Single record per item.**
2. Turn off **Combine arrays of text/numbers.**
### Custom separators
Instead of commas, you can specify a custom separator for combined values. This is useful if your data **already contains commas** and you need a different delimiter such as the pipe symbol (`|`).
Both **Combine nested object keys** and **Combine arrays of text/numbers** use the separator you specify in the **Separator** input field.
**Need More Control?**
For more advanced control over how your response is structured, you can use a [response JMESPath](/advanced-options/response-jmespath) to customize exactly how the data is extracted and transformed.