Importing a spreadsheet

A list someone keeps in a spreadsheet — an inventory, a register of extinguishers, a list of licences — becomes records: each row a record, each column a property.

Open Assets → Import a spreadsheet (also listed on the workspace’s Imports page).

1. The file

An Excel workbook (.xlsx) or a CSV file (comma, semicolon or tab separated). The first non-empty row names the columns; every row below it is a record. A workbook with several sheets asks which one. An old .xls must be saved as .xlsx or .csv first. At most 20,000 rows at once.

2. What the rows become

• Records of an existing type. ARGUS ranks the workspace’s types by how many of the columns match properties of their own, and by how close their name is to the sheet’s (or the file’s) name. The best one is proposed when it clearly fits.
• A new type, named by you, under one of the workspace’s types (Asset for equipment, Record for agreements and cases, or any other). Being under it, the new type has every property records of that kind have — description, facility, system, lifecycle, criticality, owning service, responsible, and for equipment manufacturer, model, serial number, location, inventory number, warranty — besides the columns the sheet adds.

With Use the AI ticked and an AI endpoint configured, the AI may propose an existing type or a new one and where it goes. What it proposes is checked: a type or a property it names must exist.

3. Where each column goes

Each column goes to the record’s name, its key, a property of the type (the defaults included), a new property (with the kind its values suggest: text, number, date, yes/no, a choice list of the values found), or nowhere. Columns whose names say where they go are placed by rules, which know the usual names in English and Italian (Nome, Matricola, Produttore, Ubicazione, Stato…); the AI proposes the rest. Each proposal says who made it (rule, AI, you) and how sure it is. Change any of them.

Values are read as the property expects: numbers in either decimal style (12,5 or 12.5), dates as 2021-02-03 or 03/02/2021, yes/no as yes, no, sì, x, true…, choices by their name. A location or another linked record is found by its key or its name; a person by e-mail, username or name; a group by name.

4. Check every row, then import

Check every row reads the whole file and says, row by row, what would happen: created, updated, unchanged, or why not (a choice not in the list, a number that is not one, a location not found, a required property empty, a key that belongs to a record of another type). Nothing is written. Rows with a problem are left out of the import; fix them in the file and check again.

Import creates the new type and properties, then the records — as you would by hand, through the ledger, so each value is your confirmed statement and the history says who imported it. Every record created keeps where it came from (file, sheet and row) in Source revision. The import runs in the background; it is listed on the Imports page with what it created and every row it could not import.

Updating records later

Map a column to the record’s key: a row whose key names a record of the type already here updates it (when Update records whose key is already here is ticked and you may modify records); a row without a key, or with a new one, creates a record. Importing the same sheet again changes nothing.

Fields of a restricted class you may not see cannot be written by an import, and are never overwritten by one.