Wrong long int value exported

[fengsp]

I have the code:

import tablib

data = tablib.Dataset()
data.headers = ['test_id']

a = 568883383628111872
data.append([a])
with open('test.xls', 'wb') as f:
    f.write(data.xls)

Which produces one xls file with id as 568883383628112000.

[hugovk]

Reproducible directly with newest openpyxl 3.0.0, Python 3.7.4:

from openpyxl import Workbook

wb = Workbook()

# Grab the active worksheet
ws = wb.active

# Data can be assigned directly to cells
ws["A1"] = 568883383628111872

# Save the file
wb.save("sample.xlsx")

Reported to openpyxl:

• https://bitbucket.org/openpyxl/openpyxl/issues/1355/wrong-long-int-value-exported

[claudep]

Openpyxl will not fix it as this is an Excel limitation (https://docs.microsoft.com/en-us/office/troubleshoot/excel/long-numbers-incorrectly-in-excel).
We could detect such values and force them to str to force a text format for such numbers. To be tested...

[MohammedAlkindi]

These three PRs (#637, #650, #662) all sit open, so I ran them and master on Windows 11, CPython 3.13.13, in case it helps to see how they compare. I know the thread also floated just documenting the XLS limitation, which may well be the safer call.

On a read-back of this issue's value and the 2^53 boundary, #650 and #662 come out behaviorally identical. Both turn integers above 2^53 into text cells in xls and xlsx while leaving 42 and 2^53 numeric, and both ship tests for both formats; they differ only in that #662 factors the check into a tablib.utils helper. #637 changes xls only and carries no tests. All three passed the suite here apart from one unrelated timezone-dependent failure. Worth noting either way, per the Microsoft doc linked above, 16-digit integers below 2^53 still display rounded in Excel, so the docs note would help even alongside a code fix.