Re: Fwd: Test files [WAS: Re: xlsread in Octave 3.6.4]

[PhilipNienhuis]

Markus wrote
> Am 2013-09-21 17:20, schrieb Philip Nienhuis:
> :
> <snip>
> :
> Take a look at this two screenshots (autodelete after 5 days):
> http://p.osuv.de/index.php/DBS/
> http://p.osuv.de/index.php/DV9v/

These look nice. About all I can say from those pics :-)


> :
> <snip>
> :
>> Yes: unzip.exe is simply not included in (= built with) MXE. Same goes
>> for a few other unpackers. I think it's not hard to add them to MXE,
>> probably just use bzip2 or so as a template, but I had no time yet to
>> look into it.
>> Maybe file a bug report for it.
> 
> Maybe I do this :)
> The included unzip.exe from MinGW xlsxread works fine with Windows too.

Sorry I don't follow. "MinGW xlsxread"?


> :
> <snip>
> :
>> Anyway could you please wait until I can sent you an io-1.3.4-beta (f.
>> Octave-3.7.6+), in which your unpack step has already been included
>> (i.e., for odsread)? That'll save you (and me) a bit of work.
>> ATM I have other 'distractions' - it might be after Wednesday
>> anyway... so no hurry
> 
> Yeah, no hurry :) But xlsxread is now release in Version 1.0.
> https://github.com/markuman/xlsxread

Downloaded, but I'm afraid I won't have time to look at it in detail until
Wednesday evening. I just ran a quick test (see below).


> Every example sheet in the example folder works fine for me.

Sheet 1 of testOOXML.xlsx gives me the wrong results.
In LibreOffice 4.0.1 the occupied cell range is B2:AD9 (8 rows, 29 columns).
However, xlsxread v1.0:
>> size (xlsxread ('testOOXML.xlsx', 1))
ans =

     8   104

and I see a lot of empty columns in between the rightmost 8x3 matrix.

FTR, I've requested all occupied cell ranges in your example spreadsheets
(perhaps it'll help you somewhere):
-----------------------------------------------------------------
>> xlsfinfo ('bsp4x2t2.xlsx', 'uno')
  1: Tabelle1                         (Used range ~ A1:B4)
  2: Tabelle2                         (Used range ~ A1:B4)
  3: Tabelle3                         Empty or Chart

>> xlsfinfo ('example.xlsx', 'uno')
  1: Tabelle1                         (Used range ~ A1:AMJ1234)
  2: Tabelle2                         Empty or Chart
  3: Tabelle3                         Empty or Chart

>> xlsfinfo ('excel4mac2011.xlsx', 'uno')
  1: Blatt1                           (Used range ~ A1:B5)

>> xlsfinfo ('excel2007.xlsx', 'uno')
  1: Blatt1                           (Used range ~ A1:B5)

>> xlsfinfo ('excel2010.xlsx', 'uno')
  1: Tabelle1                         (Used range ~ A1:B5)
  2: Tabelle2                         Empty or Chart
  3: Tabelle3                         Empty or Chart

>> xlsfinfo ('excel2010-emptyandcalc.xlsx', 'uno')
  1: Tabelle1                         (Used range ~ A1:E11)
  2: Tabelle2                         Empty or Chart
  3: Tabelle3                         Empty or Chart

>> xlsfinfo ('Mergedranges.xlsx', 'uno')
  1: Sheet1                           (Used range ~ J1:AMJ15)

>> xlsfinfo ('Test1.xlsx', 'uno')
  1: Tabelle1                         (Used range ~ B2:I6)
  2: Tabelle2                         Empty or Chart
  3: Tabelle3                         Empty or Chart

>> xlsfinfo ('testOOXML.xlsx', 'uno')
  1: FirstSheet                       (Used range ~ B2:AD9)
  2: ThirdSheet                       (Used range ~ A4:AMJ39)
  3: Sheet#5                          (Used range ~ F2:Q8)
  4: Sheet5                           (Used range ~ B2:F14)
  5: Sixth                            (Used range ~ A1:B3)
  6: 7th                              (Used range ~ B2:K9)
  7: Sheet_nr_3                       (Used range ~ A2:M12)

>> xlsfinfo ('Testtime.xlsx', 'uno')
  1: Tabelle1                         (Used range ~ A2:I6)
  2: Tabelle2                         Empty or Chart
  3: Tabelle3                         Empty or Chart
-----------------------------------------------------------------


> Maybe i'll going to post in the help list and ask for someone is willing 
> to test it.

Help list? (why not)
Testing is always good. But keep this in mind: testing a version which may
morph quite a bit (i.e., upon possible/anticipated absorbing into some OF
package) could be misleading. Both for the tester and for the developer.

Do you have plans to have xlsxread read strings as well? I saw that in OOXML
strings are kept in a separate "shared strings"-file; the XML nodes related
to cells containing text contents seem to have some convoluted pointers to
those shared strings. Hmmm... looks like a speed penalty...

Anyway, thank you Markus. for making a dependency-free Excel reader.

Philip




--
View this message in context: 
http://octave.1599824.n4.nabble.com/Re-xlsread-in-Octave-3-6-4-tp4657028p4657534.html
Sent from the Octave - Maintainers mailing list archive at Nabble.com.