Converter int is truncating floats

[zeycus]

Excel stores numbers as float, even numbers we enter as integers. Using the converter numbers=int truncates those floats to integers, without rounding them first. The result is that if in the cell you enter a 252, you may get a 251 when you read in python using numbers=int. That happens when the internal representation of 252 is, say, 251.999999999999423579.

I think just rounding before applying int would solve this issue nicely.

[fzumstein]

You got a point there. For now, what you can do is use a lambda expression to accomplish this: xw.Range('A1').options(numbers=lambda x: int(round(x))).value. Out of curiosity, what version of Excel/OS are you using (as we can't replicate the issue for 252)?

[zeycus]

It's Windows 8.1 Enterprise, with Microsoft Office Enterprise 2007 version 12.0.6612.1000

In the meantime I'll use what you suggested, thanks.

[ericremoreynolds]

Hi @zeycus,

The default behavior has now been changed on the develop branch, so using numbers=int will round the float before applying the int constructor. The rare (nonexistent?) cases where one doesn't want this behavior, one can specify numbers='raw int' and the int constructor will be applied without rounding.

Regards,

Eric

[zeycus]

Hi @eric,

Agree with the solution, that was awfully quick!