Fix dtype inference and session cookie handling

[Cubewise-Ronan]

• Set default dtype to str in pd.read_csv to avoid misinterpreting values like '2025' as floats.
• Fixed issue with multiple 'TM1SessionId' cookies causing CookieConflictError during session start.

File "C:\Dev\share_service\venv\lib\site-packages\TM1py\Services\RestService.py", line 789, in _start_session
    session_id = self._s.cookies.pop('TM1SessionId', None)
  File "C:\Program Files\Python310\lib\_collections_abc.py", line 954, in pop
    value = self[key]
  File "C:\Dev\share_service\venv\lib\site-packages\requests\cookies.py", line 334, in __getitem__
    return self._find_no_duplicates(name)
  File "C:\Dev\share_service\venv\lib\site-packages\requests\cookies.py", line 405, in _find_no_duplicates
    raise CookieConflictError(
requests.cookies.CookieConflictError: There are multiple cookies with name, 'TM1SessionId'

[MariusWirtz]

We don't want the value column to be str in all cases, do we?

If we set it to None instead, it will be derived depending on the data at hand

[Cubewise-Ronan]

Actually, that was the issue I was running into — setting 'Value': None causes pandas to infer the dtype, and in some cases like "2025" (which is an attribute value for an element and should be a string), it gets coerced to 2025.0 if other rows are missing values or inconsistent.

[MariusWirtz]

I understand. But we can't enforce str as the column type for the value column. Most retrieved cube data will be numeric!

In the edge cases where you load data from attribute cubes it would be easier to pass dtype=str to the execute_mdx_dataframe function.

[MariusWirtz]

Maybe this is an obsolete comment or did you want to break after the first match?