Apache OpenOffice (AOO) Bugzilla – Issue 42946
Error Bars for X values
Last modified: 2014-11-20 23:51:48 UTC
At the moment in OOo it is only possible to display errors bars in the Y-axis; this is not very useful for scientific reports, where I need both X and Y axis error bars. Similarly, I need to be able to select a range of cells to represent the errors, not just a constant + or - value or a %/standard deviation.
You know the meaning of Priority 1? I think not realy. Set to an realistic value, and also is in the planning for the new chart.
*** Issue 61024 has been marked as a duplicate of this issue. ***
IMO this is a much needed feature (still missing in 2.0.1). This might prove to be a sub-task of Issue 47264, which is to support importing MS Excel graphs with error-bars. This is due for OOo 3 (hopefully!). Also this relates to Issue 366, which requests being able to select a range of cells to specify error bars. This and X-error bars are essential for scientifc graphs. I would like to see this set to at least a milestone of OOo 3. Regards, Andrew
When I searched for xy error bars I didn't track this one down, my apologies, 61024 is a duplicate. Thanks to Andrew for pointing it out. Setting the target milestone to something tangible would be nice as it would let people know that work is progressing on it. regards Dave
Should this be assigned to the Chart component?
Corrected component, cheers.
*** Issue 64818 has been marked as a duplicate of this issue. ***
I'll take this. I'll handle the x error bars here only as the range request is already handled with Issue 366.
accepted
changing target. The target cannot be 3.0 as the fileformat does not allow to save those settings so far.
I can't believe this is being passed over in version 3.0 Didn't they rewrite the Chart module?
@phyrexicaid, well activate your colleagues to put more votes to this issue. Currently I see only 9 votes.
@iha, I have asked that those with votes on Issue #366 move their votes to this issue, that's about as much as I can do at the moment! Even though I no longer require this functionality myself, I would like to see this implemented as it would allow greater OO.o adoption (in my opinion). You have said that the target cannot be 3.0 as the fileformat does not support saving these features. Is this addressed by ODF 1.2? Regards, Dave
@phyrexicaid, thanks! :-) there are 22 votes now and I hope the count will grow still. I also think that this issue is important, but it is better to have the voice of the customers. The file format issue was addressed in the meanwhile for ODF1.2. I set the target to 3.x now.
reset to new
I wish this were incorporated into Spreadsheets. This is a pretty basic statistic tool. I don't see why it hasn't been incorporated yet. No thanks to this bug, I will be docked points off my physics homework because I don't have time to learn another stats program/syntax such as R-Project. booo.
X Error bars are a necessity for many academic studies.
I completely agree. Having X-Error bars on scatter charts would be very useful feature (in my case, I'm using it for a Timeline).
Obviously, error bars are a must , and in most scientific occasions, they occur on both Y and X axis (and vary from entry to entry). I have set up a small demonstration on how to fit recursion curves on my web page. Until recently, it could only be used in MS (needing their solver) Now, with the new non-linear solver as an add-on, it works beautifully in OpenOffice, but .. still lacks the horizontal error bars, This small worksheet uses a chi-squared , modified to include horizontal errors, the second link shows how the "blind" fitting tool of Excel is completely fooled in such cases (get the right gravity, but the wrong sign .. upwards!) http://homepages.ulb.ac.be/~frere/Labo/EffectiveVariance+EnC.xls http://homepages.ulb.ac.be/~frere/Labo/EffectiveVariance+En-ExcelWRONGC.xls
What is the status of this issue? This means I can't do work in Open Office unless I resort to drawing in the error bars manually.
Yep - that's it - this really is a sine qua non for scientific use. It strikes me as the most important feature missing from the graph component, at least from a non-corporate perspective (which /is/ still important... [grin]). Flashing pies are no compensation for the lack of something that we used to do easily by hand...
I need both X and Y Error bars to display forex market data. The y bars are for Hi-Lo price and the x bars to represent Volume data. I therefore need to be able to select a range of cells for the error values.
I was hoping to be able to point science students to Open Office, but this is a show stopper.
As a general physics teacher, I actually use Excel for graphs with both X and Y error bars. In the recent versions, new frustating bugs have been introduced in Excel and for many other reasons I would like to move to OpenOffice. But this issue #42946 has to be solved before. We are 10 teachers in our department and initiate approx. 200 students each year to the use of a spreadsheet for scientific use (error calculus, tables et graphs).
Essential for ANYTHING that has to do with science!
Open office is amazing and I'm very grateful for a free version of complete office. However this is the second or third time I have come to try and use open office over the years I have been at university and there are still no x-error bar options. This is essential to any physics student or science student in general and must be fixed for open office to grow in popularity!
I registered just to vote for this bug. At my Physics dept at uni, Openoffice calc is used in experimental analysis in pre-honors students. Having no x error bars is a big problem. The number of student/users at school/university who would benefit from this must be in the 1000s if not more.
I fully support what has been written above. I think x-error bars is a very much needed feature if one wants to spread the use of Ooffice in the scientific community. Thanks in advance!!
Are there any plans to have this implemented? That is fairy major problem for anything scientific related. It doesn't look to difficult to implement and it is over 6 years since this feature was requested for the first time. Could we at least increase priority to P2? Thanks.
Created attachment 76268 [details] A trick how to emulate X and Y error bars. This is a trick showing how one can emulate X and Y error bars in OpenOffice Calc. Not nicely looking and to very easy but works automatically. Bold cells are inputs. X+/X- Y+/Y- are values of error ranges, in this example calculated by simple formulas. All other are copies of input values arranged in such an order that chart looks like if it would have x and y values. There is even a possibility to insert regression formulas as errors will nullify itself, with one exception - the R2 (fit coefficient) will be incorrect. Hope it will be useful, Tomasz Sztejka
I don't think I've ever had to do a scatter plot without x-error bars. So many people can't use this as their spreadsheet program because of this. With y-error bars already implemented, I'm not sure I understand why this has been an issue since 2005. It seems to me a good chunk of the code would have to have been already written but for a change in axis. I guess I also don't understand why they're aren't more votes for this...shrug.
As a science student I can't live whithout X error bars.
I also just registred to vote for this issue. This is much needed for everything science, and I can't understand why this hass been an issue for 6 years. It would be great to have x-error bars implemented as soon as possible.
I was wondering who will handle openoffice bugs now that Sun isn't officially handling OO. Any chance on sending this bug list over to LibreOffice - they seem much more focused on fixing issues like the error bars.
As you might have heard OpenOffice.org is transferring to the Apache Software Foundation. The process has just started. Right now OpenOffice is in incubation state: http://incubator.apache.org/openofficeorg/index.html The new Apache OpenOffice.org project will take care of the further OpenOffice development. You are welcome to join and make things happen! Bugfixes and features made for Apache OpenOffice will reach a greater audience and in cooperation will also land in LibreOffice. In contrast everything that is only done at LibreOffice will need to be done twice because of license issues, so that would be a pity split of forces. I agree that this issue is very important and I have suggested to get it done already for a long time. I have prepared the viewing parts and file format already. Hopefully I finally will be able to work on the rest in some near future.
Yeah, it's a shame Oracle didn't see fit to turn it over to LibreOffice - I'm sure they had their reasons though. Thanks for your efforts
According to my experience this is the ONLY major bug which prevents open office from becoming popular in science. NEEDS TO BE FIXED ASAP.
Amazed that this is only marked as being of trivial importance. This is the only problem stopping me using this, and I imagine a huge number of other people - pretty much anyone who wants to use it for scientific purposes. The rest of the program is irrelevant if it can't do this. I'm resorting to buying Excel but it's a shame when Openoffice is so close to fulfilling my needs.
I wouldn't normally mention the competing open source project since it's not relevant to this bug tracker. But this is the only bug of OOo I've followed for 5+ years, and it should be interesting information for many that LibreOffice 3.6 has this feature. You'll find the feature mentioned also in the 3.6 new features and updates page. In time, I believe similar support will come to Apache OpenOffice.org as well as it continues to be developed, so let's keep this bug open for that.
I think there's a good reason for this feature not being implemented in OO. And it's the fact that spreadsheets in general should not be used for scientific data analysis. These are business and accounting tools - not scientific tools. There is a lot of specialized data analysis software - first example is the "R" project - open source, well documented and easy to learn. R's data analysis and graphing capabilities are like Heaven and Earth, compared to Calc or Excel. Moreover, Excel is known to have multiple bugs in its data analysis engine. I don't know about Calc - never tried, since I used R for the first time. If you would take my advice, drop both Calc and Excel and start using R. You'll never want to go back.
By that token Calc shouldn't exist at all since there's clearly better open source business and accounting tools. It's a general purpose grid-based mathematical manipulation tool, as such it should have x-error bars. If you're doing real science perhaps it shouldn't be used, but for high school or undergraduate university assignments you shouldn't be required to use professional tools with large learning curves.
(In reply to comment #40) > I think there's a good reason for this feature not being implemented in OO. > And it's the fact that spreadsheets in general should not be used for > scientific data analysis. This is not a fact but obviously your personal opinion. I believe X error bars make LibreOffice Calc more attractive for high school teaching in sciences, for example. The error bar in Excel or Calc should not be different from the one in R, but need less additional learning effort.
Been waiting for this to get fixed for years... But besides just the bars, it would be nice to have an error ellipse feature.