← All examples

EDJAS extraction spec for the ONS Retail Sales Index summary workbook.

The workbook is never modified; every value below is read from it. This spec shows the reference kinds a well-produced government spreadsheet needs: - single-quoted spaced sheet names ('Cover Sheet'!A1), - a named Excel Table (RevisionTriangles_Table1), resolved wherever it lives, - and a pipeline that reshapes and tidies the data (records, then round).

The specification

The whole of retail.toml, applied to the unmodified retail.xlsx.

[extract]
title     = "'Cover Sheet'!A1"
published = "'Cover Sheet'!A5"
next_due  = "'Cover Sheet'!A6"
contact   = "'Cover Sheet'!A12"
source    = "'Cover Sheet'!A15"
contents  = "['Contents'!A2:B5 | records]"
revisions = "[RevisionTriangles_Table1 | records | round 3]"

The JSON it produces

See the extracted JSON — exactly what edjas retail.xlsx retail.toml prints, which is the thing the report below is only one rendering of.

The report it produces

retail.xlsx — a flat statistics release

Retail Sales Index — Summary and Quality Tables, Retail Sales Index, Great Britain, May 2025, Office for National Statistics. Published 20 June 2025. More at ons.gov.uk.

A "Releasing statistics in spreadsheets" good-practice release: a cover sheet, a contents sheet, and data laid out as a named Excel Table (RevisionTriangles_Table1).

Downloaded from ons.gov.uk.