Hi, yes, we have a limit of 100 MB of data that the built-in xl() function is able to reference either through the grid or through a PowerQuery connection. You can split this up into smaller chunks to workaround or use PowerQuery to summarize the data before loading in Excel.

Hey @stonebig, is there a sense of data size you're looking for? It's been a while but this enhancement is still one we want to address.

100mo was looking an arbitrary low limit for no reason.

If can handle more data in jupyterlite free web than in excel paid Microsoft, I will feel punished for no reason, especially given the price difference

That would be bad, so that may be the measure of what you shall allow

I'm a big fan of pivotable, till 100 000 records, and don't like tricky cubes, when it's not sales. Maybe I'm biased

Also the datashader packages distorted my vision of "too much data", a while ago, no to mention duckdb and polars recently

imagine if soon gemini or whatever deliver a superior upload and pre-analysis power, and the result in a editable web notebook app.

Thanks for that note! We do plan to investigate increasing the limit. Will keep this open to track when it becomes available (no date yet).

My use is big etl one step, on azure data, in an easier to maintain way than M or vba

The actual problem is that I don't see where it's written about limitations and slow downs when you can't be python premium

It's a bit painful to have not it written in plain details.