Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss
custom error bars "gone wrong"! Help!
1,061 views
Skip to first unread message
lind...@officeformac.com
Feb 18, 2010, 12:04:20 PM2/18/10
to
Version: 2008
Operating System: Mac OS X 10.6 (Snow Leopard)
Processor: Intel
I�m having trouble with my custom value error bars in excel. Just got Macbook so everything is quite new, but office is a well known "buddy".
I have a bar diagram in log scale and want to add error bars. When choosing both (plus and minus) halv the diagram goes black! If I choose just plus, the bars appears, but some of them are askew making contact with error bars of the bar beside!
when choosing minus all the error lines runs to the left crossing the y-axis.
Dont have anything left to try from my heads ideas, and really need help for this!
Linda
I have a bar diagram in log scale and want to add error bars. When choosing both (plus and minus) halv the diagram goes black! If I choose just plus, the bars appears, but some of them are askew making contact with error bars of the bar beside!
when choosing minus all the error lines runs to the left crossing the y-axis.
Dont have anything left to try from my heads ideas, and really need help for this!
Linda
Mike Middleton
Feb 18, 2010, 4:40:25 PM2/18/10
to
Linda -
I suggest that the first thing you do is choose Excel
> About Excel to verify that you are using the latest Excel update:
12.2.3
- Mike
<lind...@officeformac.com> wrote in message news:59bb3...@webcrossing.JaKIaxP2ac0...
lind...@officeformac.com
Feb 19, 2010, 5:00:04 AM2/19/10
to
Hi Mike, just checked and i have the 12.1.0 version of excel. Tried to update, but no new version came as a result. Is there any other way to get the newest update?
Btw, I�ve managed to fix the positive values, but when inserting negative (operating with the box "both") the error bars of some (approx half) are askew and connect with other bars...
Btw, I�ve managed to fix the positive values, but when inserting negative (operating with the box "both") the error bars of some (approx half) are askew and connect with other bars...
lind...@officeformac.com
Feb 19, 2010, 7:47:25 AM2/19/10
to
Hmm, I updated to the newest version 12.2.3, but still the same problem. The weirdest thing is that one of the graphs worked out nicely with error bars, but the other one still have the same problem with connecting error bars...
Carl Witthoft
Feb 19, 2010, 9:22:12 AM2/19/10
to
In article <59bb3...@webcrossing.JaKIaxP2ac0>,
lind...@officeformac.com wrote:
lind...@officeformac.com wrote:
The advice to "update to most recent" was essentially useless, as
updating only helps when there's a known, repeatable bug.
In your case, it's almost certainly a misunderstanding on your part as
to what or how data in cells are mapped to error bar quantities.
Best way to solve would be for you to post your workbook somewhere that
people could download it and identify what went wrong with your graph.
Failing that, post a SMALL but representative set of data, including
the error values, here. In addition, can you please tell us exactly
which setup options you chose when selecting error bars for your chart?
--
Team EM to the rescue! http://www.team-em.com
stei...@gmail.com
Jan 14, 2016, 4:39:28 PM1/14/16
to
Error bars in Logarithmic plots end at wrong data points wenn some values are 0 or negative.
Example data:
1 0.00E+00 1.50004E-11
2 2.00E-11 2.00E-11
3 0.00E+00 4.10335E-11
4 -1.00E-12 7.64768E-11
5 0.00E+00 8.50E-10
1st column is x, 2nd y, 3rd y errors. Vertical axes is logarithmic.
In fact, only Point 2 can be plotted in a straightforward way. For the other points, the y-value is not plotable. y plus y-error is, however, valid. At that location a error bar cap is plotted, but it is connected with a diagonal line with the LAST VALID DATA POINT. This is point 2 for points 3, 4, and 5.
The error bar cap of point 1 is connected with some random point, which even moves if the zoom factor or number format in some other part of the sheet is changed.
Obviously, there is an uninitialized variable in the Excel source code, if a data point is not vaid, but excel still uses these coordinates to draw the error bar. This variable contains an arbitrary number if the invlaid point is the first point, and the coordinates of the last valid point for the further invalid points.
The expected value would be to not draw the affected error bars at all, or, more correct, draw a vertical line from the error cap to the lower end of the diagram.
I posted this as a bug with the (german) feedback tool of my office 365 (the bug also exists in office 2013), but hope for no reaction from microsoft.
I'm a university teacher for physics and electronics.
Example data:
1 0.00E+00 1.50004E-11
2 2.00E-11 2.00E-11
3 0.00E+00 4.10335E-11
4 -1.00E-12 7.64768E-11
5 0.00E+00 8.50E-10
1st column is x, 2nd y, 3rd y errors. Vertical axes is logarithmic.
In fact, only Point 2 can be plotted in a straightforward way. For the other points, the y-value is not plotable. y plus y-error is, however, valid. At that location a error bar cap is plotted, but it is connected with a diagonal line with the LAST VALID DATA POINT. This is point 2 for points 3, 4, and 5.
The error bar cap of point 1 is connected with some random point, which even moves if the zoom factor or number format in some other part of the sheet is changed.
Obviously, there is an uninitialized variable in the Excel source code, if a data point is not vaid, but excel still uses these coordinates to draw the error bar. This variable contains an arbitrary number if the invlaid point is the first point, and the coordinates of the last valid point for the further invalid points.
The expected value would be to not draw the affected error bars at all, or, more correct, draw a vertical line from the error cap to the lower end of the diagram.
I posted this as a bug with the (german) feedback tool of my office 365 (the bug also exists in office 2013), but hope for no reaction from microsoft.
I'm a university teacher for physics and electronics.
0 new messages