REGR: re-add encoding for read_excel

[twoertwein]

• closes BUG: read_excel not accepting encoding on 1.1.0 #35753
• tests added / passed
• passes black pandas
• passes git diff upstream/master -u -- "*.py" | flake8 --diff
• whatsnew entry

Needs tests!

[simonjayhawkins]

Needs tests!

There were changes to test_to_excel_output_encoding in #34464, can that test be updated to cover this.

[WillAyd]

Does this do nothing or am I misreading it? Doesn't the excel format already define how to interpret its data?

[jreback]

yeah we need a test here; I know you are trying to get one @twoertwein

[twoertwein]

Does this do nothing or am I misreading it? Doesn't the excel format already define how to interpret its data?

I think it doesn't do what I expected it to do:

It will forward the encoding to

pandas/pandas/io/excel/_base.py

Line 494 in 13940c7

parser = TextParser(

but this is the wrong place (encoding finally ends up in TextFileReader but it seems to be never used there), it should (if we have an Excel file that needs a specific encoding) go to

pandas/pandas/io/excel/_xlrd.py

Line 37 in 13940c7

return open_workbook(filepath_or_buffer)

since this is the place (or one of the places) where the file is actually opened. After that, all functions seem to operate on strings so there shouldn't be a need for an encoding form there one.

I will put this PR in draft mode until we have an excel file that needs an encoding.

Edit: closed, a future PR might go the route of forwarding the encoding to each "Excel backend" (if there is a need for that)