Make matrix of N number of (x,y) points

I want to select N random points of intersection without repetition.

1. any intersection point

2. only along perimeter

The gap between the lined horizontally and vertically are same 'D'

The number of lines may differ from that of the picture, depending on values of L1, L2, W1, W2 and D

[Adam Danz]

I have a vague understanding of what you want to do but writing solutions to vague problems is almost always a waste of time.

I find it helpful to think of coding problems as a set of inputs and a set of desired outputs that go into and out of a black box (where you code will live).

It is usually a really bad idea to start writing code for the black box without fully understanding what the inputs and expected outputs should be.

I see you have a 6-sided grid with no units or no tick values. I kind of undestand that you want to pull some coordinates from that grid but the rules for choosing coordinates is very unclear and since there are no values along the axes, I'm not even sure what those coordinates would look like. I wonder of W1, L1, etc have any meaning or need to be variables in the code. I wonder how the grid size is defined or if it changes. I wonder if the (0,0) coordinates has a defined position on the grid.

The exact inputs and examples of expected outouts need to be well defined before any code is written.

“If I had only one hour to save the world, I would spend fifty-five minutes defining the problem, and only five minutes finding the solution.”

-Albert Einstein

[Ni2]

L1, L2, W1, W2 and D are variables

L2 < L1 and W2 < W1

D is the smallest of all variables

To make the grid

1. Draw the frame

2. Draw lines vertically and horizontally inside the frame starting from bottom left corner with spacing of "D"

My requirement:

1. Select N points of intersections of lines without repetition.

2. Select N points of intersections of lines only along the perimeter without repetition.

[Adam Danz]

That's much better!

1. Is the bottom, left corner (0,0)?
2. Of your 2 requirements, do they define two different sets of output coordinates (so, there would be 2 output matrices) or are those 2 requirements applied to the same set out outputs (so, there would be 1 output matrix)?
3. Is the "missing rectangle" defined by W2 and L2 always in the upper, right corner of the rectangle formed by W1 and L1?
4. is d a multiple of W1? In other words, are the grid squares along the width of the main rectangle alway complete? In other words, if the grid units are 1, is the width always an integer?
5. Same question as #4 but for L1?
6. I see that W2 and L2 start in the middle of a grid-square. If W2 is defined as 4.5, does that mean the left side of W2 is W1 - 4.5?
7. same question as #6 but for L2?

If you provided sample values for L1, L2 W1, W2, many of these questions would be answered. Again, the best thing you could do to help us help you is to provide sample inputs and outputs (not descriptions but actual values to work with).

[Ni2]

Yes bottom corner is (0,0) for ease.

This grid can be anywhere in space though.

I want some kind of logic to store those intersection points along perimeter in a matrix A, intersection points other than perimeter in a matrix B.

Matrix C can be combination of both.

Then i wanna choose N points from A,B or C without repetition.

I am thinking to use loops for storing and random functions for choosing N random points.

[Adam Danz]

"Yes bottom corner is (0,0) for ease.; This grid can be anywhere in space though."

That sounds like "yes, but no". I'll assume the bottom, left corner is always at (0,0). If that changes it would be ease to shift the results accordingly.

What about the answers to the other questions (I added more)? If you could use the question numbers to reference the answer to those question, it would help to keep things organized.

[Ni2]

L1, L2, W1, W2 and D can be fractions and the missing rectangle is always on the top right corner.

The example above has

L1 = 6.75, L2 =3.35, W1=8, W2=4.5 and D=1

Suppose i want to choose 5 random points along perimeter.

I can expect (0, 0), (4, 0), (5, 3.5), (2,8) and (3.4, 7) as one of my output

Suppose i want to choose 5 random points not along perimeter.

I can expect (1, 1), (6, 2), (2, 5), (3, 3) and (3, 7) as one of my output

my motive is to plot some dots at that spots.

Answer 1

[Adam Danz]

Here's a solution following your definitions.

At the top of the code you can defin the input values W1 L1 W2 L2 D and the coordinates of the bottom, left of the grid.

• C is an n x 2 matrix of [x,y] coordinates for all grid points
• A is an m x 2 matrix of [x,y] coordinates for the perimeter points
• B is a p x 2 matrix of [x,y] coordinates for the inner grid points

At the end a plot is created to visually confirm that C, A and B are correct.

See inline comments.

See more text under the figure.

% Define main rectangle
W1 = 8.8;     %width
L1 = 9.8;    %Height (or length)

% Define missing rectangle in upper, right corner
W2 = 4.2; 
L2 = 6.6; 

% Define grid space
D = 1; 

% Coordinate of bottom, left of main rectangle
BL = [0,0]; %[x,y] coordinates of Bottom Left

%% Define grid points
% grid locations for main rectangle
% Later we'll remove the grid points that are defined by L2 and W2
L1pts = unique([BL(1) : D : W1, W1]); 
W1pts = unique([BL(2) : D : L1, L1]); 
% [L1W1x(n), L1W1y(n)] defines the n_th grid location of the main rectangle. 
[L1W1x, L1W1y] = meshgrid(L1pts, W1pts); 

% Grid endpoints for the missing rectangle
W2left = W1-W2; 
L2bottom = L1-L2; 

% Remove coordinates from the main rectangle that are 
% within the missing rectangle by replacing thier 
% values with NaNs
isInMissingRec = (L1W1x > W2left) & (L1W1y > L2bottom);
L1W1x(isInMissingRec) = NaN; 
L1W1y(isInMissingRec) = NaN; 

% Plot out result to make sure things make sense. 
figure()
plot(L1W1x(:), L1W1y(:), 'ko')
axis equal
grid on
set(gca, 'XTick', unique(L1W1x(~isnan(L1W1x))), 'YTick', unique(L1W1y(~isnan(L1W1y))));

%% Output intersections
% all intersections
% C is a n x 2 matrix of [x,y] coordinates for each intersection
C = unique([L1W1x(:), L1W1y(:)],'rows');
C(any(isnan(C),2),:) = []; 

% Perimeter coordinates
bottomTopPerim = cell2mat(splitapply(@(xy){[xy(1,1),min(xy(:,2)); xy(1,1),max(xy(:,2))]}, C,findgroups(C(:,1)))); 
leftRightPerim = cell2mat(splitapply(@(xy){[min(xy(:,1)), xy(1,2); max(xy(:,1)), xy(1,2)]}, C,findgroups(C(:,2)))); 
% A is a m x 2 matrix of [x,y] coordinates for each parimteter coordinate
A = unique([bottomTopPerim; leftRightPerim],'rows');

% Inner coordinates
% B is a p x 2 matrix of [x,y] coordinates for each non-perimeter coordinate
B = C(~ismember(C, A, 'rows'),:);

% Plot the results confirm it makes sense
figure()
plot(B(:,1),B(:,2),'bo', 'DisplayName', 'Inner')
hold on
plot(A(:,1),A(:,2),'rs', 'DisplayName', 'Perimeter')
plot(C(:,1),C(:,2),'k.', 'DisplayName', 'All')
axis equal
grid on
set(gca, 'XTick', unique(L1W1x(~isnan(L1W1x))), 'YTick', unique(L1W1y(~isnan(L1W1y))));
legend()

*"Innter" in legend is a type, it was fixed in the code.

Now that you've got A, B and C, it's easy to choose random coordinates.

Here' a line that chooses 5 random coordinates from C

rc = C(randsample(size(C,1), 5),:); 

[Ni2]

its not working as expected for example

W1 = 100, L1 = 150, W2 = 35, L2 = 49, D = 10

for

rc = C(randsample(size(C,1), 5),:); 

It says undefined function of variable 'randsample'

[Ni2]

i have already made an app to plot this grid. I want to plot N intersections points in the grid for

1. Along perimeter
2. Not along perimeter
3. Anywhere

Below is the screenshot of the grid which i have plotted.

[Adam Danz]

"its not working as expected for example"

once again, please add detail. We can't read your mind.

"undefined function of variable 'randsample'"

Looks like you don't have the stats & machine learning toolbox

https://www.mathworks.com/help/stats/randsample.html

You could use randi() intead but be careful because randi() samples with replacement so it could choose the same coordinate twice. Or you could look up "alternatives to randsample".

I'm curious, as my answer helped you at all?

[Adam Danz]

"i have already made an app to plot this grid. I want to plot N intersections points in the grid for"

Then why didn't you mention this in your question!??

If you've already mapped the grid, that means you already have access to all of the coordinates which was 80% of the work I ended up doing.

You already did most of the work but instead of sharing that, you required us volunteers to do repeat the work which consumes valuable time!

@Ni2, do you realize your question was asked 2 hours ago and a lot of the time since then was spent trying to understand what you want to do and then recreating what you've already done. What a waste of time.

Next time you ask a question, provide all of the information you have and spend time writing an informative question where volunteers don't have to ask dozens of follow-up questions - especially when you've already done most of the work!

[Ni2]

I am a beginner in matlab so I might not have been able to explain you things accurately. I am sorry for the inconvenience as well, I am new to this forum.

Looks like you don't have the stats & machine learning toolbox

Now the "randsample" is working normally

Thank you so much for time, please check my screenshots. The perimeter points are not in the expected place for the missing rectangle side, it is in the coordinate which is a multiple of "D". The perimeter points are at expected place for the main length and width.

below is the code to plot the grid.

i have plotted the it segment by segment, line by line

i want to make perpendicular lines in those N random points in the grid

[Ni2]

            cla (app.UIAxes,'reset')
            hold (app.UIAxes,'on')

%             l1 = 
%             l2 = 
%             w1 = 
%             w2 = 
%             d = grid spacing
%             dob = the depth of the grid (can be considered 0)
%             nr = the number of random points to required
%             lr = length of line that i need to plot in that selected random points
%             gc = position (along perimeter, not along perimeter, anywhere)

            Po = [0,0,-dob];
            Pl = [l1,0,-dob];
            Pw = [0,w1,-dob];
            Pi = [l1-l2,w1-w2,-dob];
            Pil = [l1,w1-w2,-dob];
            Piw = [l1-l2,w1,-dob];

            pts0 = [Po;Pl];
            pts1 = [Po;Pw];
            pts2 = [Pil;Pl];
            pts3 = [Piw;Pw];
            pts4 = [Pil;Pi];
            pts5 = [Piw;Pi];

            plot3(app.UIAxes,pts0(:,1),pts0(:,2),pts0(:,3),'Color','b')
            plot3(app.UIAxes,pts1(:,1),pts1(:,2),pts1(:,3),'Color','b')
            plot3(app.UIAxes,pts2(:,1),pts2(:,2),pts2(:,3),'Color','b')
            plot3(app.UIAxes,pts3(:,1),pts3(:,2),pts3(:,3),'Color','b')
            plot3(app.UIAxes,pts4(:,1),pts4(:,2),pts4(:,3),'Color','b')
            plot3(app.UIAxes,pts5(:,1),pts5(:,2),pts5(:,3),'Color','b')

            nlf = d;
            nwf = d;

            while nlf < l1-l2
                Pa = [nlf,0,-dob];
                Pb = [nlf,w1,-dob];
                pts = [Pa;Pb];
                plot3(app.UIAxes,pts(:,1),pts(:,2),pts(:,3),'Color','b')
                nlf = nlf + d;
            end

            while nwf < w1-w2
                Pa = [0,nwf,-dob];
                Pb = [l1,nwf,-dob];
                pts = [Pa;Pb];
                plot3(app.UIAxes,pts(:,1),pts(:,2),pts(:,3),'Color','b')
                nwf = nwf + d;
            end

            nlp = nlf;
            nwp = nwf;

            while nlp < l1
                Pa = [nlp,0,-dob];
                Pb = [nlp,w1-w2,-dob];
                pts = [Pa;Pb];
                plot3(app.UIAxes,pts(:,1),pts(:,2),pts(:,3),'Color','b')
                nlp = nlp + d;
            end

            while nwp < w1
                Pa = [0,nwp,-dob];
                Pb = [l1-l2,nwp,-dob];
                pts = [Pa;Pb];
                plot3(app.UIAxes,pts(:,1),pts(:,2),pts(:,3),'Color','b')
                nwp = nwp + d;
            end
            
            Pe = [1,1,0];
            Pf = [1,1,3];

            pts6 = [Pe;Pf];
            plot3(app.UIAxes,pts6(:,1),pts6(:,2),pts6(:,3),'Color','r')
            
            xlabel(app.UIAxes,'length');
            ylabel(app.UIAxes,'width');
            zlabel(app.UIAxes,'height');
            axis(app.UIAxes,'equal');
            view(app.UIAxes,[-1 -1 1]

[Adam Danz]

But none of that matters if you already have the coordinates of the grid. If you already have the coordinates of the grid you need to tell me what format they are in.

For example, are your coordinates stored in an n-by-2 matrix of [x,y] values?

The most imortant thing to learn here is to describe or supply ALL of the data you're working with in your question. We can't see your monitor and we have no idea what variables you already have. We only have the information you give us.

So you question should be been something like

I have a set of coordinates stored in the variable "grid_values" that is a 128 x 2 matrix of [x,y] values' These values form a grid in the shape of a 6-sided object as shown in the image below. I need to 1) isolate the coordinates that lay along the parimeter and 2) isolate the inter coordinates that are not on the parimeter. Then I need to choose random values from each of those sets of coordinates.

[Ni2]

I have a set of coordinates stored in the variable "grid_values" that is a 128 x 2 matrix of [x,y] values' These values form a grid in the shape of a 6-sided object as shown in the image below. I need to 1) isolate the coordinates that lay along the parimeter and 2) isolate the inter coordinates that are not on the parimeter. Then I need to choose random values from each of those sets of coordinates.

But I don't have any coordinates matrix, I have only six vertex points. Please check my codes above.