xlwings excel selection to dataframe - error with the date column - pandas

Question

I am trying to write a UDF in python/Excel using xlwings. I have time-series data in three columns in a spreadsheet of the form:
Date Hour Value
01/11/2017 1 43.1
01/11/2017 2 41.8
01/11/2017 3 38.6
01/11/2017 4 38.6
01/11/2017 5 38.6
And I want to be able to select this range, manipulate it in several ways (monthly average etc.) using pandas groupby functions then output the results back into a new spreadsheet. This code works for me:
#xw.sub
def get_df_from_range():
"""get's df"""
#make the current selection a dataframe
wb = xw.Book.caller()
df = wb.selection.options(pd.DataFrame, index = False).value
#simple check: add a sheet and print the dataframe
sht = wb.sheets.add()
sht.range('A1').options(index = False).value = df
However, as soon as I try to do any manipulation of the dataframe before printing it, I get error messages. For example:
#xw.sub
def get_df_from_range():
"""get's df"""
#make the current selection a dataframe
wb = xw.Book.caller()
df = wb.selection.options(pd.DataFrame, index = False).value
#simple manipulation task
df['Date'] = pd.to_datetime(df['Date'], format = '%d/%m/%Y')
#add a sheet and print the dataframe
sht = wb.sheets.add()
sht.range('A1').options(index = False).value = df
This gives an error message:
Run-time error '2147467259 (80004005)':
AttributeError: 'tuple' object has no attribute 'lower'
if value.lower() in _unit_map:
File
"C:\User\AppData\Local\Programs\Python\Python36-32
line 441, in f
unit = {k: f(k) for k in arg.keys()}
I thought I could debug this better by creating the same code, but not as a UDF; just by writing code in Spyder and connecting to the spreadsheet - so I would have the df variable in my variable explorer. But when I wrote almost exactly the same code, it did not give me an error message:
wb = xw.Book("my_spreadsheet.xlsm")
df = wb.selection.options(pd.DataFrame, index = False).value
df['Date'] = pd.to_datetime(df['Date'], format = '%d/%m/%Y')
I am really stuck as to why this would be. Can someone please help?
I should note, I am aware that xlwings automatically reads Excel dates as datetime64[ns] formats. That isn't the point I am trying to make. I want to do other things with the dataframe (e.g. left join it to another dataframe) and all those other tasks also fail when I try the UDF method, but work O.K. when I just connect to the spreadsheet from Spyder. I am hoping that if I can get that one "simple manipulation task" to work, then all the other tasks may also work.