read_excel does not raise error with wrong keyword arg

[lucianoviola]

For example:

pd.read_excel('file_name.xlsx',sheet='mysheet')

does not raise any error. The right parameter name is "sheetname" . If "sheet" is passed it defaults to the first sheet and no error is raised.

sheetname and sheet are very similar and easy to confuse.

One might think that it is reading the sheet passed as argument when in fact it is another one.

I would like to suggest for errors to be raised if unexpected args are passed.

[TomAugspurger]

This may be little tricky, since IIRC we pass through **kwargs to the underlying engine doing the work (though I may be wrong). Can you take a look in pandas/io/excel.py to see if there's anything we can do?

[gfyoung]

@TomAugspurger : You are correct regarding **kwargs. That is indeed quite unfortunate.

[TomAugspurger]

Hmm, we could maybe explicitly check for some common mistakes (like sheet), verify that none of the engines use them, and raise if we see them? That's fragile to the engines adding them in the future :/

[gfyoung]

@TomAugspurger : I think we should look into expanding the signature to obviate the need for **kwargs. Ultimately, we would like it to be identical to read_csv, but that's another story for another day.

[lucianoviola]

@TomAugspurger: I’m not an experienced programmer, but I could try something. What about something as simple as if parameter passed not in a list of allowed parameters, raise an error?

[gfyoung]

@lucianoviola : That sounds like a good plan. Refer to read_csv to see which arguments we can accept. Also, use the same error message that you get when you pass in an unexpected argument to a function (i.e. "[function] got an unexpected keyword argument [arg]")

[lucianoviola]

@gfyoung Ok!