How can i plot a contour plot with the information below?

I have a 57 x 3 vector where the first 2 columns represent the x and y coordinates respectively and I'd like to join the coordinates with a curve based on the value in the 3rd column.

i.e. all coordinates with a "1" in the third column will be joined by a single curve, all coordinates with a "2" in the third column will be joined by a separate curve and so on, until all 57 points have been plotted.

The vector is called "new" shown below:

0.814723686393179  0.959291425205444  1
0.905791937075619  0.547215529963803  1
0.126986816293506  0.138624442828679  1
0.913375856139019  0.149294005559057  1
0.632359246225410  0.257508254123736  1
0.0975404049994095  0.840717255983663  1
0.278498218867048  0.254282178971531  1
0.546881519204984  0.814284826068816  1
0.957506835434298  0.243524968724989  1
0.964888535199277  0.929263623187228  1
0.157613081677548  0.349983765984809  1
0.970592781760616  0.196595250431208  1
0.957166948242946  0.251083857976031  1
0.485375648722841  0.616044676146639  1
0.800280468888800  0.473288848902729  1
0.141886338627215  0.351659507062997  1
0.421761282626275  0.830828627896291  1
0.915735525189067  0.585264091152724  1
0.792207329559554  0.549723608291140  1
0.959492426392903  0.917193663829810  1
0.655740699156587  0.285839018820374  1
0.0357116785741896  0.757200229110721  1
0.849129305868777  0.753729094278495  1
0.933993247757551  0.380445846975357  1
0.678735154857774  0.567821640725221  1
0.757740130578333  0.0758542895630636  1
0.743132468124916  0.0539501186666072  1
0.392227019534168  0.530797553008973  1
0.655477890177557  0.779167230102011  2
0.171186687811562  0.934010684229183  2
0.706046088019609  0.129906208473730  2
0.0318328463774207  0.568823660872193  2
0.276922984960890  0.469390641058206  2
0.0461713906311539  0.0119020695012414  2
0.0971317812358475  0.337122644398882  2
0.823457828327293  0.162182308193243  2
0.694828622975817  0.794284540683907  2
0.317099480060861  0.311215042044805  2
0.950222048838355  0.528533135506213  2
0.0344460805029088  0.165648729499781  3
0.438744359656398  0.601981941401637  3
0.381558457093008  0.262971284540144  3
0.765516788149002  0.654079098476782  3
0.795199901137063  0.689214503140008  3
0.186872604554379  0.748151592823710  3
0.489764395788231  0.450541598502498  3
0.445586200710900  0.0838213779969326  3
0.646313010111265  0.228976968716819  4
0.709364830858073  0.913337361501670  4
0.754686681982361  0.152378018969223  4
0.276025076998578  0.825816977489547  4
0.679702676853675  0.538342435260057  5
0.655098003973841  0.996134716626886  5
0.162611735194631  0.0781755287531837  5
0.118997681558377  0.442678269775446  5
0.498364051982143  0.106652770180584  5
0.959743958516081  0.961898080855054  5

[Birdman]

See my answer.

Answer 1

[Walter Roberson]

 new = [ ...];

 [unique_labs, lab_idx] = unique(new(:,3));   %I do not rely upon them being integer or consecutive

 num_unique = length(unique_lab);
 for K = length(unique_lab) : -1 : 1
   this_label = unique_labs(K);
   this_xy = new(lab_idx == K, 1:2);
   line_handles(K) = plot(this_xy(:,1), this_xy(:,2), 'DisplayName', sprintf('group %g', this_label) );
 end
 legend( line_handles, 'show')

[Alexander Taylor]

Hi Walter,

What is the purpose of:

this_xy = new(lab_idx == K, 1:2);

I tried to run the code and got a Subscripted assignment dimension mismatch error on the following line.

[Walter Roberson]

If we knew for sure that the labels were consecutive integers starting from 1, then the code could be shortened to

for K = 1 : 5
  this_xy = new(new(:,3) == K, 1:2);
  line_handles(K) = plot(this_xy(:,1), this_xy(:,2), 'DisplayName', sprintf('group %g', K) );
  hold on
end
hold off
legend(line_handles, 'show')

and if we knew that there were no other items plotted we could shorten further to

for K = 1 : 5
   this_xy = new(new(:,3) == K, 1:2);
   plot(this_xy(:,1), this_xy(:,2));
   hold on
end
hold off
legend('group 1', 'group 2', 'group 3', 'group 4', 'group 5');

Answer 2

[Birdman]

[~,ia,~]=unique(new(:,3));
ia=[ia;size(new,1)];
for i=1:numel(ia)-1
   figure(i);
   contour(new(ia(i):ia(i+1),1:2))
end