Complaint about OpenOffice support for XLSX

[WardXmodem]

If I open an xlsx spreadsheet as downloaded from my broker, one column repeats data many times.
If I open it as a zip file, the data is _NOT_ repeated.

If I open the same spreadsheet with Libre Office, it is fine.

P.S. this forum has been very helpful in some issues.

And totally bogus for some: such as advice "OpenOffice is not a clone of Excel. Use Excel for XLSX files."
uhh, Apache open office can open XLSX files. There's no excuse for it opening them incorrectly, especially when Libre office opens the same file with no problems. seeing complaints of repeated data "for years" means no one cares to fix it. I guess development has stopped. or prioritizes new things over fixing old bugs.

SO with all the crashes and messed up formatting I get in Apache spreadsheet and writer, I'm giving up on it.
Thanks for all your help, bye bye.

[Zizi64]

Open Office v 4.2.1, Windows 10

There is not such AOO version yet.
...Are you using an archaic-old LibreOffice version? The latest LO Still version is the 7.3.5, and the latest LO Fresh version is 7.4.0 today.

Or maybe you have an unofficial version of the AOO. Where you downloaded it from?

...or maybe it is a simple typo in your signature.

[John_Ha]

I suggest that you carefully read [Tutorial] Differences between Microsoft and AOO/LO files for a description of differences, for why you should always work in, and save Writer files as .odt, Calc files as .ods, Impress files as .odp, etc, and for the problems everyone is likely to come across.

Neither AOO nor LO (nor any other program) are or claim to be clones of MS Office - they claim to be compatible with MS Office files, but not fully compatible. If you need full MS Office and MS Office files capability, buy MS Office. If you can live with missing some of the less important, more esoteric things things then AOO or LO or other program should suit you.

Apache OpenOffice Product Description

Compatible with other major office suites ...

[LO] Use documents of all kinds

LibreOffice is compatible with a wide range of document formats such as Microsoft® Word (.doc, .docx), Excel (.xls, .xlsx), PowerPoint (.ppt, .pptx) and Publisher. But LibreOffice goes much further with its native support for a modern and open standard, the Open Document Format (ODF).

[Villeroy]

OpenOffice fell into agony 10 years ago. One of the biggest advantages of OpenOffice is that it can not write OOXML formats. This way it does not help spreading this pestilence of a file format.

[John_Ha]

This way it does not help spreading this pestilence of a file format.

[dacha]

If I open an xlsx spreadsheet as downloaded from my broker, one column repeats data many times.
If I open it as a zip file, the data is _NOT_ repeated.

If I open the same spreadsheet with Libre Office, it is fine.

This terrible bug, reported many times (eg. https://bz.apache.org/ooo/show_bug.cgi?id=123306), only exists in the 4.1.X releases, it is fixed in Git.

Unfortunately we don't know when it was fixed or by what commit, so it's hard to find that commit and patch new 4.1.X releases. I will try to, when I find the time: it's essentially a "git bisect", but old OpenOffice versions are hard to compile on modern operating systems.

I don't understand all the XSLX hate, it is a widely used file format and we really should support it better. I did fix a number of bugs importing XLSX files in the past month, and these fixes will be appearing in the 4.1.14 release.

[Hagar Delest]

I don't understand all the XSLX hate, it is a widely used file format and we really should support it better.

Because it is not a true open standard. The ISO version is not the one used by MS Office by default AFAIK. The transitional format still uses features that are likely to be undocumented. This is part of the vendor lock-in policy.

LibreOffice users may be happy at some point but it may also lead to frustration when users think that LO (or any application that handles OOXML) is a good replacement of MS Office and then stumble on issues.

[John_Ha]

This terrible bug, reported many times (eg. https://bz.apache.org/ooo/show_bug.cgi?id=123306), only exists in the 4.1.X releases, it is fixed in Git.

I downloaded file showing the problem from the bug report (actually Excel_cellTextDuplication.xlsx) and it opens without duplication in my AOO 4.1.13 under W10.

Excel_cellTextDuplication.xlsx

(48.27 KiB) Downloaded 489 times

I don't understand all the XSLX hate, it is a widely used file format and we really should support it better. I did fix a number of bugs importing XLSX files in the past month, and these fixes will be appearing in the 4.1.14 release.

See [Tutorial] Differences between Microsoft and AOO/LO files for a description of differences. On the forum we are forever dealing with people who firmly believe that there is no difference between different formats and it doesn't matter which you use. I drew this to try and show the problems.
.

Differences between files.png (66.67 KiB) Viewed 13555 times

One thing which frustrates me is AOO is completely silent when opening a Microsoft file which contains tags which AOO does not recognise and does not tell the user that content in the file is not being displayed. Examples include those defining Textboxes and rsid tags. Six years ago I raised an enhancement request Issue 126927 - Enhancement Request - Notify user if any XML tags in the [.docx] file being opened are not recognised citing .docx files where Textboxes and their contents are not displayed.

[dacha]

I don't understand all the XSLX hate, it is a widely used file format and we really should support it better.

Because it is not a true open standard. The ISO version is not the one used by MS Office by default AFAIK. The transitional format still uses features that are likely to be undocumented. This is part of the vendor lock-in policy.

That's true, but that's not really the problem that we have.

Of the XSLX bugs I fixed in the last month:

Bug 126720 was caused by the XSLX specifying xl/sharedStrings.xml but actually containing xl/SharedStrings.xml, and we now work around that by trying to find files with case-insensitive names if case-sensitive fails.

Bug 127672 was caused by the XSLX rows and/or cells not containing addresses. When rows lack an address, Excel uses previous row + 1, and when a cell lacks it, previous cell's row and previous cell's column + 1. We didn't and just completely ignored the cell. Now, we do what Excel does. (Yes it's undocumented in the OOXML spec, but it wasn't difficult to discover.)

Bug 117672 was caused by _rels/.rels in the XLSX containing a path with a leading slash, when it's not supposed to have any. We now strip these unnecessary slashes, and such files open fine.

None of these problematic documents were written by Excel. They were all generated by third-party tools, but Excel just happens to allow them.

They were also reasonably easy to fix: XML is pleasant to work with, and these problems were simple and obvious.

All the negativity about OOXML convinced me that would be difficult to fix OOXML bugs, so I never even tried. Had someone told me how easy it would be, they could have been fixed 10 years earlier.

Oh and if any other developers are interested in helping to improve OOXML support, please contact me.

LibreOffice users may be happy at some point but it may also lead to frustration when users think that LO (or any application that handles OOXML) is a good replacement of MS Office and then stumble on issues.

Yes, but wouldn't it be better if OpenOffice improved and become a viable alternative to MS Office and LO?

[Zizi64]

Oh and if any other developers are interested in helping to improve OOXML support, please contact me.

This is an Users help Users forum. There are not developers here. Contact with Apache (if you want to be a developer of the AOO) or the Document Fundation, Collabora, etc... (if you want to be a develeoper of the LibreOffice)

[Hagar Delest]

... but wouldn't it be better if OpenOffice improved and become a viable alternative to MS Office and LO?

Ah, that's an interesting question.
Honestly, I don't think so. Why still have 2 open source applications both competing to rival MS Office? That's a waste of energy. It makes sense when there are significant differences and the agendas are not exactly the same. But AOO has clearly lost the battle here. Not that I'm happy about that. I have been a huge fan of AOO until a couple of years ago when I had to admit that I would be more efficient in my daily work with LO (no regret so far).
If AOO had continued to be supported by a company, then it would have made sense, one branch being corporately structured and focused (I guess) on stability, professional use, perhaps something to replace MS Office to cut costs (many did that with LO BTW), and another branch working like the bazaar model, more focused on new features, users wishes...
But today, AOO is not even able to keep up with LO. The mere issue of the files reduced to hashtags is a show stopper for me. I do know this bug needs specific conditions (not fully understood yet I think) and I have never suffered it but this kind of issue is the very reason why I switched from MS Office to OOo at that time (I was not used to backup at that time). It seems that even the integration with the OS layers (security, updates...) is now more and more difficult and requires more and more knowledge from the users to get the application just run on their machine (so many recent topics about that in the forum this last 6 months or so).
The brand power is still amazing however. I'm surprised how many people don't yet know about LibreOffice. The LO team was very successful at the beginning to gain momentum with more or less good reasons (many FUD at that time) and now that the product is clearly better than AOO, there is no real will (as far as I can see) to continue to "recruit" new users. But I may be wrong, I don't follow their team and what they do at all.

Take the exemple of The Gimp: is there any similar open source application for image edition? Would it make sense to fork it to do something similar? I don't think it will ever happen as long as there is no significant issue with the management of the project, or to do a very specific application.

[esperantisto]

But today, AOO is not even able to keep up with LO.

Still, AOO is much more stable as compared to LO, especially on large files. In my experience, of course.

Take the exemple of The Gimp: is there any similar open source application for image edition? Would it make sense to fork it to do something similar?

One answer to both questions: Glimpse

[sveld]

But today, AOO is not even able to keep up with LO.

Still, AOO is much more stable as compared to LO, especially on large files. In my experience, of course.

Mnn that is not my expirence. I'm always on the most "fresh" LO and I hardly see it crashing, then the point I run Fresh is so I can find issues<G>
LO did a lot to the codebase to make it more stable and keep it that way. Beside a lot of code cleanups, leveraging Coverty Scan to keep track of issues in the code which worked out pretty well. https://scan.coverity.com/projects/libreoffice

Take the exemple of The Gimp: is there any similar open source application for image edition? Would it make sense to fork it to do something similar?

One answer to both questions: Glimpse

Oh yes, that's an excellent example: ambition does not magically clone a community

[MrProgrammer]

Thanks for all your help, bye bye.

Ward really did say bye bye this week.
Ward Christensen, BBS inventor and architect of our online age, dies at age 78