Bug 393213

Summary:	If (0,0) is present in a spreedsheet, fitting fails
Product:	[Applications] LabPlot2	Reporter:	urcindalo
Component:	backend	Assignee:	Stefan Gerlach <stefan.gerlach>
Status:	RESOLVED FIXED
Severity:	normal	CC:	stefan.gerlach
Priority:	NOR
Version:	unspecified
Target Milestone:	---
Platform:	Other
OS:	Linux
Latest Commit:	https://commits.kde.org/labplot/bc977b884f2fe68e2b949a2093332e58434a2ff5	Version Fixed In:	2.5

Description urcindalo 2018-04-17 08:12:41 UTC

Hi!
I'm checking out Labplot2 2.5.0 (last sync and installation, 2018/04/16).
When you try to fit your data to a custom non-linear equation, like Michaelis-Menten's [Vm*x/(Km+x)], and your spreedsheet contains the initial (0,0) pair of data, you get this failure:
"Fit status: Iteration is no making progress towards solution".
When you remove (0,0), fitting proceeds.

Comment 1 Stefan Gerlach 2018-04-22 10:44:30 UTC

Git commit bc977b884f2fe68e2b949a2093332e58434a2ff5 by Stefan Gerlach.
Committed on 22/04/2018 at 10:42.
Pushed by sgerlach into branch 'master'.

1. fixed bug in finite difference calculation for custom fitting models
2. add test for bug 393213
FIXED-IN: 2.5

M  +6    -4    src/backend/worksheet/plots/cartesian/XYFitCurve.cpp
M  +58   -0    tests/analysis/fit/FitTest.cpp
M  +3    -1    tests/analysis/fit/FitTest.h

https://commits.kde.org/labplot/bc977b884f2fe68e2b949a2093332e58434a2ff5

Comment 2 urcindalo 2018-04-23 12:19:43 UTC

(In reply to Stefan Gerlach from comment #1)
> Git commit bc977b884f2fe68e2b949a2093332e58434a2ff5 by Stefan Gerlach.
> Committed on 22/04/2018 at 10:42.
> Pushed by sgerlach into branch 'master'.
> 
> 1. fixed bug in finite difference calculation for custom fitting models
> 2. add test for bug 393213
> FIXED-IN: 2.5
> 
> M  +6    -4    src/backend/worksheet/plots/cartesian/XYFitCurve.cpp
> M  +58   -0    tests/analysis/fit/FitTest.cpp
> M  +3    -1    tests/analysis/fit/FitTest.h
> 
> https://commits.kde.org/labplot/bc977b884f2fe68e2b949a2093332e58434a2ff5

Thanks!