calc problem

Hi, before I log this as a bug I would like to know whether anyone has seen it before/potential fix identified.

I normally use Fedora linux and I download (.XLS) share tables from a financial newspaper here in Oz. Yesterday I inadvertently clicked on the wrong table by accident and tried to open it. Calc locked up! No problem with the files I normally look at, Calc opened them fine.

So I tried again today and waited for an error - "LibreOffice 4.0.1.2 General input/output error". So I hopped over to the windows PC and tried LO portable. Same error. The error was repeated whether I tried to open directly from the website or if I saved to disk and then attempted to open the file. MS Excel has no problems.

The file in question is accessible using this link:

http://www.afr.com/share_tables/

Just click on the xls link next to "Warrant Market"

Any comments?

Cheers

Hello Tim,

Just click on the xls link next to "Warrant Market"

Any comments?

Yes, it locked up here too, using v3.5.4.2 in Debian, at about 50%
through loading. Top showed 100% CPU usage. After a minute or so, I
killed the process.

I confirm it does not open with LO 4.0.1.2 (Linux Mint Maya 64). Also, Calligra Sheets had problems. And it opened (I think correctly) with Gnumeric - the version from the Ubuntu Precise repository.

I just went to "100 Leading Industrial Stocks" on that page and LO 3.6 opensuse opened straight up. "Warrant Market" doesent but locks LO. The csv opens ok.

Steve

Thanks Jay and Brad. I'll raise a bug report

Ironically that site has language along the lines of "you will need MS Excel to open these files". I was going to write to correct them on this. Best keep my mouth shut now :frowning:

Cheers

It opens OK in the Mac, LO 4.0.1.2 version.
114 Rows by 13 columns.

Hi :slight_smile:
It opens fine in Gnumeric but not LibreOffice 4.0.0 on Ubuntu 12.04.  So far it only seems to open on a Mac. 
Regards from
Tom :slight_smile:

Tim:

Thexls files are XML files with an xls file extension. Amaya, a xmleditor, bloats to nearly a 1 gig of memory, when this file is opened.

TomW

To follow up on the xls file being an xml file, I tried to open several xls files in the same xml editor with complete failure. The test files do open correctly in Calc.

How they saving an xml file as xls? Is someone changing the file extensions?

Hi Michael, looks like this one didn't make it to the list. And it is very interesting. In light of all the XML reports I think the first step is to ask the site why XML docs are masquerading as XLS. It would be good to know the version of MSO you are using so I can quote this to them.

FYI I am using Office 2007 and the file opens without this message,

Thanks to all for your feedback so far.

Cheers

It is not an .xlsx file. .xlsx files are a kind of Zip package, similar to how .ods is a special Zip package.

Rename it to .xml.

Try opening it, in LibreOffice, as "Microsoft Excel 2003 XML."

Also, Excel should be able to open it as an XML file unless it has been damaged.

There are some data extraction programs that produce these as .xls when they are an Excel form of XML. This may be one of those situations.

- Dennis

Hi Tim,

Since you use Fedora Linux, maybe you can try gnumeric as a temporary
alternative.

Hope it helps and regards,

Michael

Is there a reason why the .xls file version is preferred to the .csv
version from the afr.com site?
nvsoar

It was not a matter of preferences. It was because calc could not open a .XLS file. On further investigation it was found that it is not an XLS file hence the problem

Cheers

Hi :slight_smile:
So it meant that instead of discovering a bug with LibreOffice it was a discovered problem with the 3rd party's website and their strange way of setting up their downloads.

The question was whether reporting a bug against LibreOffice might improve LibreOffice's handling of Xls files but the answer to that is just "No" and that solves the problem.
Regards from
Tom :slight_smile:

Hi :slight_smile:
So it meant that instead of discovering a bug with LibreOffice it was a discovered problem with the 3rd party's website and their strange way of setting up their downloads.

The question was whether reporting a bug against LibreOffice might improve LibreOffice's handling of Xls files but the answer to that is just "No" and that solves the problem.
Regards from
Tom :slight_smile:

IMHO it not necessarily a bug but certainly a feature request.

Hi :slight_smile:
Yes, i agree.  Not handling a weird sort of Xls is not necessarily a bug.  Native formats might be more of an issue.  Since it's an ultra-weird format and not even an Xls it's definitely not an LO bug.

Btw many apols for being rude earlier.  I'm a bit of a pillock sometimes.  Sorry chap. 
Regards from
Tom :slight_smile:

Jay Lozier 提到:

Hi :slight_smile:
So it meant that instead of discovering a bug with LibreOffice it was
a discovered problem with the 3rd party's website and their strange
way of setting up their downloads.

The question was whether reporting a bug against LibreOffice might
improve LibreOffice's handling of Xls files but the answer to that is
just "No" and that solves the problem.
Regards from
Tom :slight_smile:

IMHO it not necessarily a bug but certainly a feature request.

Because I can open the downloaded xls and csv file with gnumeric without
problem, maybe something in LibreOffice is needed to be checked.

Regards from

Michael

Hi :slight_smile:
You are welcome to post it as a bug.  One of the drop-down menus lets you choose "enhancement request" or something if you want
https://wiki.documentfoundation.org/BugReport

Regards from
Tom :slight_smile:

Tim:

I finally had enough free time(very rare(short leash)), with access to Excel '03, to work with the file. I started by opening it in Excel and truncating rows of data. I would then try to open this truncated file with Calc (latest portable version). I started with large chunks (about 1k row). Once Calc would open the file (4k rows removed), I started splitting the difference between 3k and 4k rows removed. I kept getting the Data I/O failure until I had removed 3617rows from the end. I then repeated the process by removing rows of data from the beginning of the file. Again the file would open in Calc with 3618 rows removed from the beginning of the file. In another block of free time, I removed rows from the middle of the filewith similar results.

Items of note:

The original file used 7200 rows in Excel.

I neverremoved any rows from the Column labels on up.

The pass/fail for opening the file in Calc was unstable when near the 3617row removal.

When cancelling the Data I/O error message, it would take nearly the same amount of time for the message to clearas it did to appear.

The original file structure had 191k lines. Row/cell tags were quite often on their own line from the data.

Once opened in Excel and saved the file structure was automatically changed to Excel Spreadsheet XML format. This seemed to have no effect on Calc's ability to open the files.

When Calc was able to open the file, two empty rows above the column labels were missing that were there when opened in Excel.

I did not have enough time to try truncating data using a text editor on the original file.

TomW