Clipboard functions don't support Unicode on Windows

[jcjf]

Recent versions of pyperclip (pandas.util.clipboard) use CF_UNICODETEXT getting data from and putting data to the WIndows Clipboard.

[jreback]

what needs to change? what is the probemetic statement? want to do a pull-request?

pls pd.show_versions()

[jcjf]

Currently, git master for pandas is based on pyperclip 1.3, the latest version is 1.5.7. The Unicode features for Windows were added very recently. I'd be tempted to replace the code pandas.util.clipboard with https://github.com/asweigart/pyperclip/blob/master/pyperclip/__init__.py. I don't know how to compile and test pandas on Windows (I've always used Christoph Gohlke's binaries).

My use case at the office was that I was trying to load clipboard data from Excel that was in Spanish. The current code in pandas.util.clipboard couldn't handle this. I am at home now but I think a UnicodeDecodeError was thrown - it should be easy to reproduce this issue with non-ASCII clipboard data.

[jreback]

ahh, ok. Yes we can just update in master to the new version. Though would like some tests for validation.

[jcjf]

Since pyperclip is now a PyPI package, is it better to make it an optional dependency?

[jreback]

we could update the pyperclip package. Don't want to add another specific dep (e.g. that you might have to download).