[Libreoffice-bugs] [Bug 127624] Add Excel-compliant version of COUNTIF

[bugzilla-daemon at bugs.documentfoundation.org]

https://bugs.documentfoundation.org/show_bug.cgi?id=127624

Xisco Faulí <xiscofauli at libreoffice.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Severity|normal                      |enhancement
                 CC|                            |xiscofauli at libreoffice.org
     Ever confirmed|0                           |1
            Summary|COUNTIF irregular result    |Add Excel-compliant version
                   |                            |of COUNTIF
             Status|UNCONFIRMED                 |NEW

--- Comment #13 from Xisco Faulí <xiscofauli at libreoffice.org> ---
(In reply to Winfried Donkers from comment #10)
> Created attachment 154277 [details]
> simplified use cases
> 
> (In reply to Julien Nabet from comment #5)
> 
> I can explain what happens: 
> -if the cell to be compared and the criterion are of the same type, the
> comparison result is true;
> -if the cell to be compared is of type numeric and the criterion is of type
> text, the comparison result is true, i.e. the cell to be compared is
> evaluated as text; 
> -if the cell to be compared is of type text and the criterion is of type
> numeric, the comparison result is false.
> (see uploaded document, rows 3..6)
> 
> COUNTIF is defined in
> http://docs.oasis-open.org/office/v1.2/os/OpenDocument-v1.2-os-part2.
> html#COUNTIF and the criterion is defined in
> http://docs.oasis-open.org/office/v1.2/os/OpenDocument-v1.2-os-part2.
> html#Criterion. In these paragraphs I cannot find explicit statements
> accepting or rejecting the behaviour as described above. 
> The '=' operator, which can be used in the criterion, is defined in
> http://docs.oasis-open.org/office/v1.2/os/OpenDocument-v1.2-os-part2.
> html#Infix_Operator_EQ. Here the use of two different type should produce
> false as result. If we use '=' in the criterion Calc does not comply with
> this. (see uploaded document, rows 8..11)
> I cannot say which is fully compliant with the ODFF standard and which is
> not, probably Eike can tell. 
> 
> Excel produces a result true for all 8 combinations used in the uploaded
> document. However, if COUNTIF in Calc is compliant with the ODFF standard,
> COUNTIF is not at fault, but we may need an Excel-compliant version of
> COUNTIF.

I agree. Let's turn this into an enhancement then. Thanks for the great
analysis

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/libreoffice-bugs/attachments/20190925/6bb199ae/attachment.html>