https://bugs.documentfoundation.org/show_bug.cgi?id=106796
Winfried Donkers <[email protected]> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |ASSIGNED
CC| |[email protected]
Assignee|[email protected] |winfrieddonkers@libreoffice
|desktop.org |.org
--- Comment #2 from Winfried Donkers <[email protected]> ---
(In reply to raal from comment #0)
> Description:
> [...]
> 2. The skipping of rows when only one of the two is missing is illegal as
> far as I am concerned. Excel says "If actual_range and expected_range have a
> different number of data points, CHISQ.TEST returns the #N/A error value."
> and ODFF has constraints "ROWS(A) == ROWS(E)" and "COLUMNS(A) ==
> COLUMNS(E)". So this looks like a bug in Excel and behaviour that should
> _not_ be copied into Calc.
>
> [...]
In my text in Comment 1 I missed an important phrase in ODFF1.2:
'For an empty element or an element of type Text or Boolean in A the element at
the corresponding position of E is ignored, and vice versa.'.
This does not explain what is to be done with the degrees of freedom should a
complete row or column be empty. Excel keeps the degrees of freedom unchanged,
i.e. solely based on the range dimensions. I will do that too for Calc.
There is one -undocumented- exception in Excel: when calling CHITEST/CHISQ.TEST
with ranges that are completely empty, the result should be 1, but Excel
returns #DIV/0!.
I propose that Calc returns an error (Illegal Argument) and that a change
request is made for ODFF: A- nor E-ranges may be empty, there must be at least
1 pair in the data set.
--
You are receiving this mail because:
You are the assignee for the bug._______________________________________________
Libreoffice-bugs mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/libreoffice-bugs
