Apache OpenOffice (AOO) Bugzilla – Issue 40578
warn the user when choosing "logarithmic regression" with data series containing zero (0)
Last modified: 2013-09-25 20:41:48 UTC
Hi! I'm a biology student and have lots of diagrams that need a logarithmic scale and regression to be evaluable. So I often get to the problem in OpenOffice that the logarithmic parts of Chart are not working as they should. When I need a logarithmic regression and click the button for it, nothing happens. Linear expression works, but that doesn't help me. If you do a linear regression when you have one logarithmic scale (let's say x-scale) you get a curve (which is actually correct, though a logarithmic regression with a logarithmic scale should be a line too). But when I try "logarithmic regression" I don't get anything. Without that function Chart is not usable for natural science analysis at all. I can send you an example-file if this bug is unreproduceable, though I think it isn't.
Created attachment 21455 [details] A logarithmic regression is enabled but nothing to see
duplicate *** This issue has been marked as a duplicate of 366 ***
closed as duplicate
With due respect but this issue is NOT AT ALL a duplicate of 366. There may be a certain connection between error bars and a regression line and actually both functions are to enable at the same register card _but_ it is neverless a complete different function. I do not deny that I was once annoyed about the issue 366 as well but I found this bug more important. In Excel it is even possible to show the calculated function of the regression curve! Greetings Phil
according to http://www.openoffice.org/servlets/ReadMsg?list=releases&msgNo=7690 this issue will be set to OOoLater
Your data must not start with "0", then this will work. You can show the data of the curve using the LINEST and LOGEST functions
cloph, you are right. That's it. Zero is not defined for a logarithmic scale. There should be either an ignorance or an error message in this part of the program, shouldn't there?
changed summary - original summary: "logarithmic regression doesn't work" confirming, setting keywords, defaulting prio, defaulting TM
I take this.
accepted
Changed Target to 2.x
back to later
reset to new