OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.

 


Help: OASIS Mailing Lists Help | MarkMail Help

office message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]


Subject: [OASIS Issue Tracker] Commented: (OFFICE-2170) From Office-662-'NumberSequence' is not properly defined



    [ http://tools.oasis-open.org/issues/browse/OFFICE-2170?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=19077#action_19077 ] 

David Wheeler  commented on OFFICE-2170:
----------------------------------------

Array semantics are in the array section.  If there's a problem with them, I think they should be fixed in that section; is that inadequate?  If so, suggestions?

As far as what happens in a rectangular range, the quick answer is that the rectangular area is transformed into a sequence.  Different implementations do it differently, though.  E.G., Excel 2007 seems to do "row at a time", while OpenOffice.org 3.2.0 does it a "column at a time".  Given that difference, and the fact that no one seems to notice, I would leave row-at-a-time vs. column-at-a-time as an implementation-defined option.  See below for more detail.

I intend to add a section noting what Sequences are, see OFFICE-1187 for more about that.

== DETAILS ==

As promised, here is more detail about row-at-a-time vs. column-at-a-time processing for sequences.

Microsoft Excel 2007 seems to convert a cuboid reference to a sequence this way ("row at a time"):
Start at the initial sheet, initial row, initial column.  Process each *row* in turn, from smallest to largest column value (e.g., A1, B1, C1).
Once a row has been completely processed, go on to the next row.  When all rows on a sheet have been processed, go to the
next sheet and start over if there is a next sheet.

OpenOffice.org  seems to convert a cuboid reference to a sequence this way ("column at a time"):
Start at the initial sheet, initial row, initial column.  Process each *column* in turn, from smallest to largest row value (e.g., A1, A2, A3).
Once a column has been completely processed, go on to the next column.  When all columns on a sheet have been processed, go to the
next sheet and start over if there is a next sheet.

Presumably, if there's a concatenation of cuboids, they would process the first concatenation this way, then add the next concatenated cuboid to the sequence, and so on.
I didn't test that.

I determined this by creating a test Sheet1 with:
1	2	3	
4	5	#DIV/0!	#N/A
7	#N/A	9	

And test Sheet2 with:
1	2	3
#N/A	5	6

All starting at A1.

In my tests, I presume that if an error is encountered, the FIRST error encountered is returned.
That's true for operators, and I don't know why it would be different for ranges.
If that's false, please let me know!

In Excel 2007:
=SUM(A1:C3) gives #DIV/0 - so it does a row, then the next row, etc.
=SUM(A2:D2) gives #DIV/0 - so inside a row, it does "left to right".
  I say that in quotes because it RTL languages, like Arabic or Hebrew, that
  will actually *display* right to left, but it'll still compute inside a row from
  lowest to highest column value.
=SUM(Sheet1:Sheet2!A1:C3) gives #DIV/0, so it does a whole sheet before
 going on to the next sheet.

In OpenOffice.org 3.2.0:
=SUM(A1:C3) gives =NA(), so it computes by COLUMNS.
I also checked, and inside a column it computes lowest to highest value row.



> From Office-662- 'NumberSequence' is not properly defined
> ---------------------------------------------------------
>
>                 Key: OFFICE-2170
>                 URL: http://tools.oasis-open.org/issues/browse/OFFICE-2170
>             Project: OASIS Open Document Format for Office Applications (OpenDocument) TC
>          Issue Type: Bug
>          Components: OpenFormula
>            Reporter: Eric Patterson
>            Assignee: David Wheeler 
>            Priority: Minor
>             Fix For: ODF 1.2
>
>
> 'NumberSequence' is not properly defined - it does not specify 
> what happens in an array, nor what happens for a rectangular 
> range (more than one row / column). There is no section actually 
> defining it - merely a 'Conversion to...' section.

-- 
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: http://tools.oasis-open.org/issues/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira

        


[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]