Error in Global Fit

Share icon

Thu, 11/15/2012 - 03:24 am


I started having errors when running the global fit package proposed in Igor.
I am not completely sure, but it seems that the problem started appearing when using the latest version (6.22A).

The error is in the else part of the following:

Wave/T ListWave = root:Packages:NewGlobalFit:NewGF_DataSetListWave

if (WaveExists(w))
ListWave[row][NewGF_DSList_YWaveCol][0] = NameOfWave(w)
ListWave[row][NewGF_DSList_YWaveCol][1] = GetWavesDataFolder(w, 2)
else
ListWave[row][NewGF_DSList_YWaveCol][0] = "" // this allows us to clear the data set from a row
ListWave[row][NewGF_DSList_YWaveCol][1] = ""

And when opening a blank Igor file and a global fit, the waves of the last global fit are appearing in the control panel. Is that normal?
I unistalled and reinstalled Igor on my computer but this did not help.

Thanks!
I see that code, but in simple tests I can't find an action that actually causes the code to be run.

What is the error message?

What did you do to provoke the error?

Is it data dependent?
Quote:
And when opening a blank Igor file and a global fit, the waves of the last global fit are appearing in the control panel. Is that normal?

No, unless you are opening an Igor experiment file (.pxp extension) in which you saved a copy of the control panel.

John Weeks
WaveMetrics, Inc.
support@wavemetrics.com

November 15, 2012 at 09:25 am - Permalink


I have the following errors when trying to open GF panel in a .pxp file where there is already some data loaded:

- Error when adding a new data set in the GF panel:

a wave write error: "Index out of range for new wave : "NewGF_DataSetListWave"."

- Then error when trying to link the parameters:

a wave write error: "Couldn't find the given dimension item label %backColors in wave "NewGF_MainCoefListSelWave"."

Hope this can clarify!

November 16, 2012 at 02:09 am - Permalink


Hi,

GF worked well for some time with version 6.3. I now have 6.32 A and get again some errors in the GF. The debugger (shown below) says: Listbox error: "Size out of range" and the problem seems to be at line "ListBox NewGF_Tab0CoefList".
I can still run the GF after clicking on the green arrow, so it does not seem to be a major error.
Do you have any idea?
Thanks!

// ControlInfo/W=NewGlobalFitPanel NewGF_TabControl
// switch(V_value)
// case 0:
Variable listTop, listHeight, dataSetsListWidth, dataSetsListRight, coefsListleft, coefsListWidth

CalcListSizes(listTop, listHeight, dataSetsListWidth, dataSetsListRight, coefsListleft, coefsListWidth)
ListBox NewGF_DataSetsList, win=NewGlobalFitPanel#Tab0ContentPanel, pos={dataSetsListRight-dataSetsListWidth, listTop}, size={dataSetsListWidth, listHeight}
ListBox NewGF_Tab0CoefList, win=NewGlobalFitPanel#Tab0ContentPanel, pos={coefsListleft, listTop}, size={coefsListWidth, listHeight}
Groupbox NewGF_Tab0ListDragLine,win=NewGlobalFitPanel#Tab0ContentPanel, pos={dataSetsListRight+NewGF_Tab0ListGrout/2, listTop},size={1, listHeight}
// break;
// case 1:
GetWindow NewGlobalFitPanel#Tab1ContentPanel wsizeDC
Variable Width = (V_right - V_left)
Variable Height = (V_bottom - V_top)
ListBox NewGF_CoefControlList, win=NewGlobalFitPanel#Tab1ContentPanel,size={width-NewGF_CoefListWidthMargin, height-NewGF_CoefListHeightMargin}
// break;
// endswitch
end

July 31, 2013 at 06:46 am - Permalink

I'm sorry that you're having trouble with Global Fit. Can you tell me what you do when you see the error? It seems fairly likely that you are doing something in an order that I didn't anticipate, and the code doesn't check for sanity.

John Weeks
WaveMetrics, Inc.
support@wavemetrics.com

July 31, 2013 at 12:53 pm - Permalink

There seems to be a conflict with my routine procedure used to load data. Indeed, if I load data by hand, I don't have any errors in the GF. I cannot remember when this has started to happen; actually except for the error ListBox size out of range, the fit seems to work...

Then I tried with an experiment containing my normal loading data routine and deleted the whole procedure from it, leaving a blank procedure, as well as deleting all stored waves, leaving exclusively the ones I needed for the GF. This gives again the same error message.

Unloading the GF, then killing all waves in "NewGlobalFit" and "NewGlobalFit_StoredSetup" and then reloading the GF package does not help either.

Do you have any idea where could be the problem? Thanks!

August 2, 2013 at 03:06 am - Permalink

Please send a copy of the experiment file that triggers the error to support@wavemetrics.com. If your procedures are stored in files external to the experiment file, I will need those, too. The easiest thing is to adopt them into the experiment. Hold down the shift key and pull down the File menu. Select Adopt All, make sure Adopt User Procedure Files is checked, and click Adopt. Then do Save As to save a copy and send it to me.

John Weeks
WaveMetrics, Inc.
support@wavemetrics.com

August 2, 2013 at 08:58 am - Permalink

A.M. sent me a copy of the experiment file in question. It turned out that the experiment contained a saved size for the Global Fit control panel that was too small. I think we may have added some items in the lower part of the panel, making it bigger. When the saved size is set, the space remaining for the list boxes in the top part is too small.

The saved information can be deleted from the wave root:Packages:WindowCoordinates:W_windowCoordinates.

John Weeks
WaveMetrics, Inc.
support@wavemetrics.com

August 5, 2013 at 12:39 pm - Permalink