ENH: na_rep='=na()' as default value

[mhooreman]

Feature Type

• Adding new functionality to pandas

• Changing existing functionality in pandas

• Removing existing functionality in pandas

Problem Description

Hello,

Excel has a formula for N/A: =na(). If is shown as #N/A.

Why not using this as default value for na_rep instead of ''?

So: df.to_excel('foobar.xlsx', na_rep='=na()')

Thanks a lot,
Michaël

Feature Description

def df.to_excel(..., na_rep='=na()', ...)

Alternative Solutions

N/A

Additional Context

No response

[rhshadrach]

Running

df = pd.DataFrame({'a': [1, 1, 2], 'b': [3, 4, 5], 'c': ["=B2+B3", "=B2+C2", "=na()"]})
df.to_excel('test.xlsx')

in LibreOffice on Ubuntu, I get

When I upload the same sheet to Google Docs, the values appear as expected

Is this just an issue with LibreOffice? I'd be hesitant to have this as a default without it working in the major Excel programs.

[mhooreman]

Running

df = pd.DataFrame({'a': [1, 1, 2], 'b': [3, 4, 5], 'c': ["=B2+B3", "=B2+C2", "=na()"]})
df.to_excel('test.xlsx')

...

Is this just an issue with LibreOffice? I'd be hesitant to have this as a default without it working in the major Excel programs.

Surprising, as =na() is a documented Excel fomula. It's most probably a missing feature in LibreOffice.

Have you tried using directly the =na() formula in LibreOffice?

But ... well ... the major Excel program is ... Excel. I don't want to debate about spreadsheets programs, but Excel file is the file format for Excel, which is supported by other programs ... who are natively using the open document fomat.

[rhshadrach]

It's most probably a missing feature in LibreOffice.

Have you tried using directly the =na() formula in LibreOffice?

=na() works when writing it in LibreOffice

But ... well ... the major Excel program is ... Excel. I don't want to debate about spreadsheets programs, but Excel file is the file format for Excel, which is supported by other programs ... who are natively using the open document fomat.

If this change is known to degrade the behavior for some users, then I'm rather negative toward it. Perhaps it's a bug in LibreOffice that can be fixed and then we can implement.

[rhshadrach]

Using Ctrl-Shift-F9 in LibreOffice resolves the values. Perhaps this is not done by default for security concerns?

A much larger issue is that this breaks round tripping. If you do not open the Excel file and evaluate the formulas, they resolve to 0 upon reading.

df = pd.DataFrame({'a': [1, 1, 2], 'b': [3, 4, 5], 'c': ["=B2+B3", "=B2+C2", "=na()"]})
df.to_excel('test.xlsx')
print(pd.read_excel('test.xlsx'))
#    Unnamed: 0  a  b  c
# 0           0  1  3  0
# 1           1  1  4  0
# 2           2  2  5  0

If I instead open the written Excel file and evaluate formulas, I get the expected results.

#    Unnamed: 0  a  b    c
# 0           0  1  3  2.0
# 1           1  1  4  4.0
# 2           2  2  5  NaN

Unless there is a way to appropriately roundtrip the NaN here without having to open the file, I don't think =na() should be the default.

[mhooreman]

Finally, is it rejected ?
Sorry ... I'm sorting my pending issues...

[rhshadrach]

That is how I would vote, the number one drawback in my mind is with round-tripping. However the idea does have some merit, and I do not like closing such issues without opinions from others. I'll mark it as a closing candidate for now.

[mhooreman]

Thanks