Selective plotting from .csv file

Hi,

I've 2 sets of results from another software. Say Design1 and Design2. In each Design I've a set of X and Y values to plot. For Design1 the length is 1mm and for design2 the length is 2mm. So now, is it possible, If I type in 1mm it has to select design1 and plot the values under it. I've more than 200 Design results with x and y values under each. So I need to write a program so that If i type in a value, it'll select which design it is and plot the value.

Could someone pls help a little.

Answer 1

[Michael Haderlein]

Will the design number always be equal to the length? Or might Design163 correspond to a length of 250 mm?

If the numbers are always the same, you can simply open the file by

n=7;
XY=dlmread(sprintf('%s\\Design%d.txt',currentpath,n));

Alternatively, you can of course also ask for the number with the inputdlg.

If the design number doesn't directly link to the length, you need to store the correlation additionally. Say, Design1-4 link to 1,2,5,10 mm, the correlation would be

c=[1 52 5 10];
l=5; %length in mm
n=find(c==l);

then read as shown above.

Plotting of course by

plot(XY(:,1),XY(:,2))

Answer 2

[Ian]

Hi Michael,

No the design number is not equal to the length. Actually there are more parameters (around 20) to select each design, for eg: length= 5mm , width= 7mm, thickness = 2mm...etc. So If I type in these values for 20 parameters, It has to choose which design and plot. If I need to store each corelation separately, it would be like 200 Designs and 20 parametes each, almost 4000 corelations.

Should I write each correlation manually?

[Michael Haderlein]

Where is the correlation stored? In the Design files? Or how do you know it?

[Ian]

Its just an output file. I have attached an example with 6 parameters for each design. These 6 values I wuold type in.

[Michael Haderlein]

I don't quite understand the matter now. So the content of your file is

Design 1,,,Design 2,,,,,,,,,,
Moment,Angle,,Moment,Angle,,Parameters,Length,Width,Thickness,temperature,x,y,z
0.05,2,,0.5,10,,Design1,5,21.8,8,50.87,0.45,0.0028,0.2
0.09,5,,0.55,16,,Design2,8.2,10,2,21.65,0.015,0.012,0.1
0.12,9,,0.61,24,,,,,,,,,
0.18,15,,1.25,29,,,,,,,,,
0.4,23,,2.2,35,,,,,,,,,
0.57,30,,3,41,,,,,,,,,
0.71,42,,3.4,50,,,,,,,,,
0.76,51,,4.2,57,,,,,,,,,
0.82,62,,5.7,61,,,,,,,,,
0.95,70,,6.5,68,,,,,,,,,

Now if you type in "5,21.8,8,50.87,0.45,0.0028,0.2" (7 parameters) you want to have Design1, if you instead type in "8.2,10,2,21.65,0.015,0.012,0.1", you want to have Design2? In case you decide for Design1, is the data you want to plot the entire first column (0.05-0.95) vs. the second column (2-70)?

Are all designs saved in one file or are there more files with the other 198 designs?

[Ian]

Hi Michael,

Perfect. That is exactly what i want. All the design data(1-200) are saved in the same file.

[Michael Haderlein]

Ok, but then the format is really strange. Just to be sure, the first two lines which have numbers (so lines 3/4) have both the design definition AND x/y data stored?

[Ian]

Actually the x/y datas are the results we get from another software in a .csv file. And yes the design definition also gets stored in the same file. But It shouldn't be a problem. We can shift those data's to a new .csv file or maybe we can change the format in the same file. The problem is to link these two.

[Michael Haderlein]

Ok, in this case you can use the following code:

fid=fopen('example.csv');
firstline=fgetl(fid);
ncolumns=1+sum(firstline==',');
ndesigns=length(regexp(firstline,'[^,]*'));
nparams=ncolumns-3*ndesigns-1;
data=textscan(fid,[repmat('%s%s%*s',1,ndesigns) '%*s' repmat('%s',1,nparams)],'delimiter',',','collectoutput',1);
fclose(fid);
values=str2double(reshape(data{1}(2:end,1:2*ndesigns),size(data{1},1)-1,2,ndesigns));
parameters=str2double(data{1}(2:ndesigns+1,2*ndesigns+1:end));

inputparams=str2double(inputdlg({'Length';'Width';'Thickness';'Temperature';'x';'y';'z'}))';
design=find(all(bsxfun(@eq,parameters,inputparams),2));

outputdata=values(:,:,design);
plot(outputdata(:,1),outputdata(:,2))

I hope this will also work with your larger file.

[Ian]

Hi Michael, Thank you very much. It works super fine. Thanks a lot.

Answer 3

[Ian]

Hi Michael, Thanks again for the code. Is it possible to use this code in the form of a GUI. I made a small gui using GUIDE, but not sure how to include this in it.

[Michael Haderlein]

Ok, my problem is that I never use GUIDE, so my answer might be odd for someone who is used to GUIDE (in case I want a GUI, I do it by hand). However, what you can do is

 hedit=get(handles.uipanel1,'children');
 inputparams=str2double(get(hedit([7,4,1,8,6,5,3]),'string'))';

The reordering of the objects ([7,4,...]) is necessary as the order of the objects is different from what it's supposed to be. Also, there's no box property, so I just left it. Anyway, if you replace the inputparams=... line by the two lines above, it should work.

[Ian]

Hi Michael, Should i change this in the GUI code or the code you sent. If I change it in your code, It shows the following error,

    Undefined variable "handles" or function "handles.uipanel1".

    Error in Interface1 (line 14)
    hedit=get(handles.uipanel1,'Int1');

[Michael Haderlein]

Just in your GUIDE m code. You have the line where you define inputparams=... in the puhsbutton1_Callback. Replace this line by the ones I have posted.