Google Sheets add-on

The Grunddata add-on brings the Grunddata workflow into Google Sheets. You load a table into a sheet, edit it like any other spreadsheet, and upload your changes as a commit.

It is the Google Workspace equivalent of the Excel plugin, and both talk to the same API with the same validation.

Installing

1. Install the Grunddata add-on from the Google Workspace Marketplace. Your Google Workspace administrator may need to approve it for your domain.

2. Open any Google Sheet.

3. Go to Extensions → Grunddata → Open Grunddata. The sidebar opens on the right.

Tip

Keep the sidebar open while you work. It shows the current sheet’s context, the number of pending edits, and the domain value picker.

Signing in

Click Sign in with Microsoft in the sidebar. Authentication uses your organisation’s Microsoft Entra ID account — the same account you use for the Grunddata web app.

A browser window opens for the sign-in, then returns you to the sheet. Your name appears at the top of the sidebar, along with a Sign out button.

Sign-in is per Google account

The add-on stores your session against your Google user. If you switch Google accounts you will need to sign in to Microsoft again.

Loading a table

Once signed in, the sidebar shows three dropdowns in sequence:

1. Environment — the Grunddata environment you want to work against. If your organisation has only one, select it and move on.

2. Model — the model containing your table.

3. Table — the table to load.

4. Click Load table into sheet.

The add-on creates a sheet named after the table and fills it with the table’s active records. Columns follow the attribute order you configured.

System columns are how the sheet knows what it is editing

The sheet contains hidden system columns holding each row’s source_id and related metadata. They are hidden for a reason.

• Never delete, sort away, or overwrite them. Doing so orphans the row: the upload will treat it as a brand new record instead of an update to the existing one.
• Use Show system columns in the sidebar if you genuinely need to see them, and hide them again afterwards.

One table per sheet

Each loaded table becomes its own sheet within the spreadsheet, so you can keep several tables side by side in one file and switch between them with the tabs.

Editing

Edit cells exactly as you would in any spreadsheet. The add-on tracks which rows you have touched.

Colour coding

The sidebar legend explains the row and cell colours:

Colour	Meaning
Orange	Edited locally, not yet uploaded
Green	Uploaded successfully
Red	Upload failed — see the upload_error column
Yellow	Skipped — not sent in this upload
Light blue cells	Domain attribute cells — use the picker

The sidebar also shows a badge with the number of pending (dirty) rows.

Adding new rows

Type into the first empty row below the data. Leave the system columns empty — that is how the add-on knows this is an insert rather than an update.

Do not sort or cut and paste rows

Sorting the sheet, or cutting and pasting whole rows, can separate a row’s visible values from its hidden identifiers. Filter instead of sorting, and edit values in place.

Picking domain values

Cells belonging to a domain attribute are shaded light blue. They must contain a value that exists in the referenced table.

1. Click the domain cell you want to fill.

2. The Domain section appears in the sidebar.

3. Type part of the code or name in the search box and click Search.

4. Choose the value from the results dropdown.

The cell is filled with the readable label (SEK - Swedish krona). On upload the add-on resolves that label back to the underlying source_id, which is what actually gets stored. See Identifiers.

Copy down after picking once

For a column where many rows share the same value, pick it once with the picker and then copy the cell down. The label resolves on upload just the same.

Do not hand-type domain values

A hand-typed label with a trailing space or a slightly different name will fail to resolve and the row will be rejected. Use the picker.

Uploading changes

1. Type a commit message describing what you changed.

2. Click Upload changes.

3. The add-on sends only the rows you edited, along with the message.

4. Rows turn green on success and red on failure.

The commit message is your audit trail

Add Q3 counterparties from the onboarding list will make sense to someone in six months. update will not. See Commits and versions.

When rows fail

Failed rows turn red and an upload_error column is written with the message returned by the server for that row — the same message you would see in the web app.

The valid rows in the same upload are still processed. Fix the red rows and upload again.

Common causes are listed in Record audit and errors.

Deleting rows

Select the rows in the sheet and click Delete selected rows.

Note

Deletes are soft. The record stops appearing in listings, downloads and OData, but its history remains. See Commits and versions.

Other sidebar actions

Action	What it does
Refresh context	Re-reads the current sheet’s context, dirty row count and selected domain cell — without touching your data
Refresh table	Re-fetches the table from the server
Autofit columns	Sizes the columns to their content
Show / hide system columns	Toggles visibility of the hidden identifier columns
Sign out	Ends the Microsoft session

Refresh table discards local edits

Refresh table replaces the sheet contents with what is on the server. Any edits you have not uploaded are lost. Upload first, or use Refresh context if you only wanted to resynchronise the sidebar.

Troubleshooting

Symptom	Likely cause and fix
Sidebar says the sheet is not a Grunddata table session	You are on a sheet that was not loaded by the add-on. Switch to the loaded sheet, or load a table.
A row uploads as new instead of updating	Its system columns were cleared or overwritten. Refresh the table and redo the edit.
Domain value will not resolve	The label was typed rather than picked, or the value does not exist in the domain table. Use the picker; add the value to the lookup if it is genuinely missing.
A table is missing from the dropdown	You may not have access to its model, or you have the wrong environment selected.
Dates rejected on upload	Format the column as text in yyyy-MM-dd (or yyyy-MM-dd HH:mm for datetime).
Sign-in loop	Sign out, close the sidebar, reopen it and sign in again.

Good practice

Work in slices on large tables

Loading a table with hundreds of thousands of rows into a spreadsheet is slow for everyone. For bulk maintenance of very large tables, use the API; for reading, use the OData feed.

Upload often, in logical batches

Small, frequent uploads with clear messages produce a history you can actually read — and a failed upload only sends you back a few rows instead of a day’s work.