Skip to main content
Query tabular data from a CSV or XLSX resource using the data.gouv.fr Tabular API. No download required.

​
Parameters

​
question
string
required
Description of what data you’re looking for.This parameter provides context for the query. It doesn’t affect filtering but helps document the query purpose.
​
resource_id
string
required
The unique identifier of the resource to query.This ID can be obtained from the list_dataset_resources or get_resource_info tool results.
​
page
integer
default:1
Page number for pagination.Use this to retrieve additional pages of data. Each page contains up to page_size rows.
​
page_size
integer
default:20
Number of rows to return per page.Minimum: 1, Maximum: 200. Start with the default (20) to preview data structure, then increase if needed.
​
filter_column
string
Name of the column to filter on.Must be used together with filter_value. Column names are case-sensitive.
​
filter_value
string
Value to filter for in the specified column.Must be used together with filter_column. The comparison operator is specified by filter_operator.
​
filter_operator
string
default:"exact"
Comparison operator for filtering.Supported operators:
  • exact - Exact match
  • contains - Contains substring
  • less - Less than or equal to
  • greater - Greater than or equal to
  • strictly_less - Strictly less than
  • strictly_greater - Strictly greater than
​
sort_column
string
Name of the column to sort by.Column names are case-sensitive. Use with sort_direction to control sort order.
​
sort_direction
string
default:"asc"
Sort direction for results.Supported values:
  • asc - Ascending order
  • desc - Descending order

​
Returns

Returns a formatted text response containing:
  • Resource title and ID
  • Dataset title and ID
  • Question/query context
  • Applied filters and sorting (if any)
  • Total row count from Tabular API
  • Number of rows retrieved
  • Total pages and page size
  • Column names
  • All retrieved data rows with full content (values truncated to 100 characters if longer)
  • Pagination information (if more pages available)
  • For large datasets (>1000 rows): Warning about using download_and_parse_resource for full analysis

​
Usage notes

This tool only works for CSV and XLSX files that are available via the Tabular API (CSV ≤ 100 MB, XLSX ≤ 12.5 MB). Use get_resource_info to check availability first.

​
Query strategy

  1. Preview structure: Start with default page_size=20 to understand the data structure and column names
  2. Small datasets (<500 rows): Increase page_size or paginate through all pages
  3. Large datasets (>1000 rows): Use download_and_parse_resource for comprehensive analysis instead of paginating

​
Filtering and sorting

You can combine filters and sorting: 
filter_column="city"
filter_value="Paris"
filter_operator="exact"
sort_column="population"
sort_direction="desc"
This queries rows where city exactly matches “Paris”, sorted by population in descending order.

​
Pagination tips

  • Check the “Total pages” in the response to know how many pages are available
  • The tool will indicate if more pages are available with a next page suggestion
  • For datasets >1000 rows, consider downloading the full file instead of paginating
For large datasets requiring full analysis, use download_and_parse_resource instead of paginating through many pages with this tool. The Tabular API is optimized for exploratory queries, not bulk data retrieval.

​
Error handling

The tool handles the following error cases:
  • Invalid filter_operator: Returns list of supported operators
  • Invalid sort_direction: Returns list of supported directions
  • Resource not available: Returns error if resource is not in Tabular API
  • HTTP errors: Returns status code and error details
  • Empty results: Returns warning if no rows match filters

​
Example workflows

​
Preview data structure

query_resource_data(
  question="Show me a sample of the data",
  resource_id="abc123",
  page_size=20
)

​
Filter and sort

query_resource_data(
  question="Find Paris population data sorted by year",
  resource_id="abc123",
  filter_column="city",
  filter_value="Paris",
  filter_operator="exact",
  sort_column="year",
  sort_direction="desc",
  page_size=50
)

​
Paginate through results

1. query_resource_data(resource_id="abc123", page=1, page_size=100)
2. query_resource_data(resource_id="abc123", page=2, page_size=100)
3. Continue until no more pages available

Build docs developers (and LLMs) love

Get started for freeTalk to us