Spreadsheets

How a spreadsheet is read

• STYD finds the tables in each sheet (a header row and the rows below it) and loads them into a private space for your workspace. You see them under the Uploaded files connection.
• A short AI check looks at the headers and a few rows of each table to confirm what is a table and what is a layout. The AI Guide of Uploaded files describes each table.
• Notes, titles, dashboards, text boxes, chart titles and cell comments are not tables. They are added as passages you can search, and a source names the sheet and the area.
• A CSV can be UTF-8 or Windows-1252 and use a comma, semicolon, tab or pipe. A JSON file must be a list of flat records, or an object that holds one such list.

Things that surprise people

• Formulas are not recalculated. STYD uses the value saved in the file. A file produced by a script without saved values can show empty cells: open it in Excel and save it first.
• Total rows are kept out of the table. Rows that add up other rows (SUM and SUBTOTAL) are not counted as data, so they are not counted twice.
• Hidden sheets, rows and columns are read. The file's details say so. Remove what you do not want STYD to read before you add the file.
• Pictures are not read. A chart or a picture in a workbook is listed, but the text inside it is not read.
• Links to other workbooks are not opened. Pivot tables are read as the plain cells they show.
• Macros are refused. Save a plain .xlsx.

Limits

A spreadsheet counts as one file and is counted in rows, not pages. One file can have up to 300,000 rows, and one table up to 200 columns. How many rows a workspace can keep and load each month depends on the plan: see Limits.

Replacing a spreadsheet

Add a file with the same name and confirm. STYD loads the new version and keeps the table names, so questions about the table keep working. Deleting the file removes its tables.