---
title: Spreadsheets | Developer Documentation
description: How Parse handles Excel, CSV, and other spreadsheet inputs, with one result page per sheet, options for sub-tables, formulas, and hidden sheets, and the DCF template example.
---

Parse accepts spreadsheet files as input and returns each sheet as one page of the result. A sheet becomes a markdown table that keeps its row and column positions, so a worksheet with a title block, an assumptions block, and two forecast tables stacked vertically comes back as one grid you can read top to bottom. A few `input_options.spreadsheet` controls change how messy sheets are split up and how formulas are evaluated.

## When to use it

- Financial models, budgets, and forecasts where the cell values matter and the layout is not a single clean table.
- Workbooks with several logical tables stacked in one sheet, separated by blank rows.
- Files that were edited but never recalculated, so cached formula values are stale.
- Exports and data dumps in `csv`, `tsv`, or `ods` that you want in the same markdown form as everything else.

Supported inputs include `xlsx`, `xls`, `csv`, `tsv`, `numbers`, and `ods`, plus `xlsm`, `xlsb`, `dif`, `sylk`, `dbf`, and others listed under [Supported document types](../../../general/supported_document_types/). The Excel example uses the `agentic` tier, a strong default for table-heavy documents.

## Options

| Option                                                          | Type    | Default | What it does                                                                                                                                                                            |
| --------------------------------------------------------------- | ------- | ------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| `input_options.spreadsheet.detect_sub_tables_in_sheets`         | boolean | unset   | Find and extract several tables within one sheet instead of merging them. Useful when data regions are separated by blank rows or columns.                                              |
| `input_options.spreadsheet.force_formula_computation_in_sheets` | boolean | unset   | Recompute formula cells instead of using cached values. Enable for files edited but never recalculated, or templates with placeholder values. Can slow parsing on formula-heavy sheets. |
| `input_options.spreadsheet.include_hidden_sheets`               | boolean | unset   | Parse hidden sheets as well as visible ones. By default, hidden sheets are skipped.                                                                                                     |
| `output_options.save_output_pdf`                                | boolean | unset   | A PDF copy of the parsed document; not produced for spreadsheet, plain-text, or audio inputs.                                                                                           |

These options apply to spreadsheet inputs and are ignored for other file types. The reverse direction, writing tables found in a PDF out as an XLSX workbook, is `output_options.tables_as_spreadsheet`; see [Tables](../tables/).

## Example

Parse a workbook and print the second sheet:

```
from llama_cloud import LlamaCloud


client = LlamaCloud()  # reads LLAMA_CLOUD_API_KEY from the environment


result = client.parsing.parse(
    file_id="FILE_ID",  # an .xlsx uploaded with client.files.create(file=..., purpose="parse")
    tier="agentic",
    version="latest",
    input_options={
        "spreadsheet": {
            "detect_sub_tables_in_sheets": True,
            "force_formula_computation_in_sheets": True,
        }
    },
    expand=["markdown"],
)


# One result page per sheet, in workbook order; page_number 2 is the second sheet
second_sheet = next(p for p in result.markdown.pages if p.success and p.page_number == 2)
print(second_sheet.markdown)
```

## What you get

The markdown for a sheet is a pipe table whose columns match the sheet’s columns, with empty cells left empty. This excerpt is the second sheet of the DCF template from the [Excel example](../../examples/parse_excel_sheets/), which stacks an instructions block, an assumptions block, and a forecast table in one sheet:

```
|Discounted Cash Flow Excel Template||||||||||||
|-|-|-|-|-|-|-|-|-|-|-|-|
|Assumptions||||||||||||
|Tax Rate|20%|||||||||||
|Discount Rate|15%|||||||||||
|||||||||||||
||5 Year Weighted Moving Average|||||||||||
||Year 1|Year 2|Year 3|Year 4|Year 5|Year 6|Year 7|Year 8|Year 9|Year 10|Terminal Value|
|Pre-tax income|50,000.00|55,000.00|45,000.00|52,000.00|60,000.00|||||||
|Net Cash Flow|20,000.00|27,000.00|23,000.00|29,600.00|35,000.00|29,093.33|29,817.78|30,177.48|30,469.23|30,379.74|287,188.00|
|Discounting Factor||||||0.8696|0.7561|0.6575|0.5718|0.4972|0.4972|
```

Because the whole sheet is one grid, joining the pages into a prompt gives a model enough context to answer questions such as which years’ income taxes equal 20% of pre-tax income. The Excel example finishes with exactly that: a small question-answering step over the parsed markdown.

## See also

- [Parse and analyze Excel spreadsheets](../../examples/parse_excel_sheets/): the full example in Python, TypeScript, Go, Java, and the CLI, with the optional question-answering step
- [Configuring Parse: spreadsheets](../../guides/configuring-parse/#spreadsheets-xlsx-csv)
- [Supported document types](../../../general/supported_document_types/) for every spreadsheet format Parse accepts
- [Tables](../tables/) for table output from PDFs and scans, including XLSX export