BUG: to_excel() cuts off list of values when creating Excel file about pandas HOT 4 CLOSED

Thanks for the report! In the example you provided, print(len(a)) outputs 1425 elements and print(len(d)) also outputs 1425 so nothing is being cutoff for me, could you double check if this example works or not?

from pandas.

Excel has a hard limit on the size of the contents in a single cell: up to 32,767 characters:

https://github.com/pandas-dev/pandas/blob/main/pandas/io/excel/_base.py#L1360

The example provided actually works, because the length of a - converted to a string - is right at the 32,767 limit. It breaks if you add another value, in which case the user will get the warning in the code above.

In that case, Excel will cut off the list right at the limit, and it seems @josemariagarcia95 suggests that the Pandas Excel writer should gracefully truncate the written data to fit the maximum Excel cell size.

However, it seems like a pathological use case of the Excel writer, because it's not normal to try to store an entire large data series in a single Excel cell; the point of Excel is to save structured data, so the more reasonable way to use it would be to write the series as a proper Excel column, by replacing the line:

b.loc[0] = [a]

with

b['a'] = a

which works perfectly.

from pandas.

Can be closed as a duplicate of #56954 This is a known Excel limitation so round-tripping is not possible here. This is going to warn in pandas 3.0 (see the linked PR of that issue).

from pandas.

I wasn't aware of the Excel limitation, sorry for the disturbance and thank you all for the help!

@chaoyihu that structure comes from a DB in my company where each row holds information about a physical signal and the column "Value" contains a huge array with all the values that were read for that signal during a short period of time (I think it's a couple of seconds but it's a high-frequency signal). I was hoping to export it to Excel so non-technies colleagues could open certain snapshots I provide them with using Excel (since the alternative was to handle them really big CSV files that they would open with a regular Notebook app and that would traumatize them instantly).

Again, sorry for the misunderstanding. Thanks!

from pandas.