> For the complete documentation index, see [llms.txt](https://docs.diaflow.io/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.diaflow.io/workflow-builder/nodes/built-in-tools/retrieve-data-from-spreadsheet.md).

# Retrieve Data from Spreadsheet

## Retrieve Data from Spreadsheet

### What does it do?

The **Retrieve Data from Spreadsheet** node lets you ask questions about a spreadsheet and get back the matching data. Think of it like handing a spreadsheet to an analyst and saying, “Show me the top products,” without sorting or filtering the file yourself.

#### Smart Features

* **Reads common spreadsheet files**: Works with `.csv`, `.xls`, and `.xlsx` files.
* **Understands plain-English questions**: In **Use AI** mode, you can ask for totals, top items, or filtered results in everyday language.
* **Prepares the sheet for you**: If you do not define the column structure yourself, Diaflow figures it out automatically.
* **Cleans extra spaces**: It removes accidental spaces in text cells before running your question.
* **Avoids re-importing the same file**: If the same file is already prepared, the node reuses it instead of loading it again.

<figure><img src="/files/0zLT7f9d6zGY6g1k0G2f" alt=""><figcaption></figcaption></figure>

### Real-World Business Value

* Check which products, regions, or teams generated the highest revenue in a sales file.
* Pull only overdue invoices or pending orders from an uploaded report.
* Let operations teams ask spreadsheet questions in plain English instead of editing filters by hand.

### Step-by-Step Setup

* In the **Data** field, select the spreadsheet file from an earlier step using **@**.
* Make sure the file is a `.csv`, `.xls`, or `.xlsx` file.
* In **Action**, keep **Query data from table** selected.
* In **SQL Generate Method**, choose **Use AI** for the easiest setup.
* In **Provider**, leave the default option unless your team has told you to use a different one.
* In the **Input** field, type a plain-English question such as `Show the top 5 products by total revenue`.
* In **Schema**, leave it empty unless your team already prepared a custom table structure for this file.
* Only use **Manually enter SQL** if you already know exactly which custom query you want to run.
* Run the workflow, then open the output to review the returned rows and totals.

<figure><img src="/files/zkwZqsXuJYgewnEWr4zv" alt=""><figcaption></figcaption></figure>

### The Transformation: Before & After

**Before**

```
Spreadsheet file: sales_q1_2026.xlsx

Rows inside the file:
Date       | Region | Product   | Revenue
2026-01-04 | APAC   | Widget A  | 12000
2026-01-09 | EMEA   | Widget B  | 8700
2026-01-12 | APAC   | Widget A  | 14300

Question:
Show total revenue by product
```

**After**

```json
[
  {
    "Product": "Widget A",
    "total_revenue": 26300
  },
  {
    "Product": "Widget B",
    "total_revenue": 8700
  }
]
```

<figure><img src="/files/Cq5xSBTdCyLqfnyrTifg" alt=""><figcaption></figcaption></figure>

### Tips &  Warnings for First-Timers

* Use a real spreadsheet file, not a web link pasted into the question field.
* If you send multiple files, only the first file is used. Split them into separate steps if needed.
* Keep filenames clear and word-based, such as `regional-sales-europe.xlsx`. Avoid filenames that only differ by numbers or symbols.
* Very large spreadsheets can load slowly or fail. Start with a smaller file when testing.
* If your spreadsheet changes, re-upload it with a new file name or path before running again. This helps avoid older results.
* Date columns may behave like plain text. Ask simple date questions first, then confirm the output.
* In **Use AI** mode, the result size is kept short. If you need a longer list, narrow your question or use a more specific filter.
* If a column name in your file is unusual or misspelled, the result may fail. Check the spreadsheet headers first.
* Use **Manually enter SQL** only for advanced use. Most teams should stay with **Use AI**.

### Need help?

* Learn the basics in [How a node works](/getting-started/lets-start-with-the-basics/how-a-node-works.md)
* Build the full flow in [Create a workflow](/workflow-builder/create-a-workflow.md)
* Browse related nodes in [Component List](/workflow-builder/component-list.md)