Each issue below follows the same shape: what it means, likely causes, and how to fix it.
“Saved datasets or variables couldn’t be loaded” on a brand-new workbook
What it means: the Add-in couldn’t read stored datasets or variables when the workbook opened.
Likely causes: a brand-new workbook has none yet, or Excel was slow to hand over storage.
How to fix it: an empty Datasets and Variables list on a new workbook is normal, and the Add-in retries on its own. If the message sticks around, close and reopen the Astrato panel. Still stuck? Note the time and your Excel version and send them to your support contact.
#NO_DATA
What it means: there’s no data to show for that formula.
Likely causes:
The dataset doesn’t exist (check the name).
There’s no cached data for it yet.
You’re offline and nothing’s cached.
Its refresh mode is Refresh manual and it hasn’t been refreshed.
How to fix it: refresh it from the panel.
#LOGIN_REQUIRED
What it means: the dataset needs a connection but can’t find usable details.
Likely causes: the connection is missing, renamed, or incomplete.
How to fix it: check that the connection still exists, that its name matches the one saved on the dataset, and that it has a URL, tenant ID, and token.
#VALUE!
What it means: Excel couldn’t evaluate the formula.
Likely causes:
A parameter or variable value that isn’t valid.
A parameter name that doesn’t exist on the measure (check the spelling — it has to match exactly).
An unsupported value.
How to fix it: correct the value or name and let the formula recalculate.
The table won’t spill (#SPILL!)
What it means: The ASTRATO.TABLE formula needs empty room to fill in and can’t get it.
Likely causes: cells below or to the right of the formula aren’t empty.
How to fix it: clear the cells below and to the right of the formula and try again. For bigger datasets, put the table on its own sheet.
I changed connection and my semantic layer vanished
What it means: switching connection cleared the dataset’s selection.
Likely causes: this is expected — data models don’t carry across environments.
How to fix it: pick the semantic layer again for the new connection.
I don’t see a new dimension or field yet
What it means: the Add-in is still showing what it already knew about the semantic layer.
Likely causes: it shows cached fields first, then quietly picks up anything new; offline, it can’t check at all.
How to fix it: open the dataset editor, wait a moment, and search again. If you’re offline, reconnect first.
“The connection is valid, but the data can’t be accessed”
What it means: your sign-in is fine, but you don’t have access to that particular data right now.
Likely causes: you haven’t been granted access to that data.
How to fix it: use Check access in Astrato / Open in Astrato from the message, confirm you can open the same data there, and request access from the data owner or an admin. No need to reconnect.
“Connection needs attention”
What it means: your saved sign-in is no longer valid.
Likely causes: an expired session or an out-of-date token.
How to fix it:
Sign in with Astrato — open the connection in Settings, sign in again, and retry.
Personal Access Token — open it in Settings, update the token, save, and retry.
A shared workbook shows cached data but won’t refresh
What it means: the recipient can read your cached numbers but can’t pull fresh data.
Likely causes: connections belong to each person, not the file.
How to fix it: the recipient signs in with their own connection (see Connect to Astrato), then refreshes on their own access.
Log level
Log level controls how much behind-the-scenes detail the Add-in records about what it’s doing. It’s a troubleshooting tool — you’ll only need it if Astrato support asks you to capture detail while looking into an issue. The levels go from silent to most detailed, and each one includes everything above it.
Note: Leave Log level set to Off for everyday use. Turning it up doesn’t change your data or how the Add-in works — it just records more diagnostic information.
When support asks you to use it:
Open Settings > Advanced and set Log level to the level they suggest (usually Info or Log).
Reproduce the problem so it gets recorded.
Share what’s captured with support.
Set Log level back to Off when you’re done.