Selecting parts of gridded data- I Need Help. I want to know it's possible or not?

Hey all hope you have a good time,

I have a (1 x 1171) cell which includes tables. Each table has a specific latitude and longitude (columns). All 1171 tables latetudes and longitudes together represent gridded precipitation data over the whole country in the 0.5 x 0.5 degree (longitude x latitude).

I also have some points data (in the form of a table with columns of longitude, latitude, and precipitation).

I want to delete each grid in my cell that doesn't have at least one of my points inside it.

For example: If red dots are my points (in the table that I mentioned above), I want to keep each grid that has at least one red dot inside it (I would like to keep grid if more than one point is in it). So from 130 grids in this figure below I want to keep only nine grids that have points. So if my cell was 130x336 now it is just 9x336.

I attach my cell and my points.

Thank you so much for your valuable advice.

[darova]

Can you please upload data for grid the same format as points? Can't read it

[darova]

It would be great if you will make a simple drawing of the result you want to see

[BN]

Dear darova,

First of all, I want to appreciate you for reading my question. I upload grid again in this comment with the same format as points.

It would be great if you will make a simple drawing of the result you want to see

I have gridded_datamat.mat when I plot them I see a 0.5 x 0.5 degree (lon x lat) pixels (grids) on map. On the other hand, I have points.mat (scatter) that when I add them on this grids, some of the grids has owned one, two, or more point inside and some grids haven't any point inside itself. I want to delete these grids that don't match with any of the points from gridded_datamat.mat.

The output is something like the gridded_datamat.mat that I uploaded here but with lower row than 1171 because undoubtedly many of grids haven't a chance to own at least one the points (I want to save each grid that at least has one point or higher).

I truly appreciate you. Please don't hesitate to contact me if something I've written isn't clear.

Best Regards

Answer 1

[darova]

Here is my achievement

clc,clear

load points.mat
load grid_datamat.mat

p = table2array(Points);
g = table2array(grid_data(:,2:3));

x = floor( (p(:,2)-24.25)/0.5 )*0.5+24.25;
y = floor( (p(:,3)-44.25)/0.5 )*0.5+44.25;

plot(p(:,2),p(:,3),'.r')
hold on
plot([x x+0.5 x+0.5 x x]',[y y y+0.5 y+0.5 y]','.-b','markersize',15)
plot(g(:,2),g(:,1),'.g')
hold off

[BN]

Dear darova,

Thank you so much. You are awesome. I think my problem can be solved now. Just one thing. All green points are one centroid (centre of a pixel, when the pixel is in 0.5 x 0.5-degree resolution (square)). Now if red points fall into one of this pixel (0.5 x 0.5 box around the centre of each green centroid), I want to keep this specific pixel (centroid latitude and longitude from gridded_datamat.mat points). But in the code above for example, sometimes four green points Being involved, I want to select only one green point that has this redpoint (based on 0.5 x 0.5 degree that I mentioned that exist for each green point).

Also how I can selects this latitude and longitude from gridded_datamat.mat? (deleting centroids that haven't any red point in this file) and have new_gridded_datamat.mat

I want to tell you I edite the code above to this:

p = table2array(Points);
g = table2array(grid_data(:,2:3));

x = floor( (p(:,3)-24.25)/0.5 )*0.5+24.25; %change 2 to 3
y = floor( (p(:,2)-44.25)/0.5 )*0.5+44.25; % change 3 to 2
plot(p(:,3),p(:,2),'.r') %change 3 to 2 and 2 to 3
hold on
plot([x x+0.5 x+0.5 x x]',[y y y+0.5 y+0.5 y]','.-b','markersize',15)
plot(g(:,1),g(:,2),'.g') %change 1 to 2 and 2 to 1
hold off

Thank you so much

[darova]

• All green points are one centroid (centre of a pixel, when the pixel is in 0.5 x 0.5-degree resolution (square))

OK. I got it. Just change these lines (remove 0.25)

x = floor( (p(:,3)-24)/0.5 )*0.5+24; % change 2 to 3
y = floor( (p(:,2)-44)/0.5 )*0.5+44; % change 3 to 2

THe result

• Also how I can selects this latitude and longitude from gridded_datamat.mat?

You can use pdist2 to find appropriate indices

xc = x + 0.25;
yc = y + 0.25;
D = pdist2([xc yc],g);          % every possible distances
[i,j] = find(D<0.01);           % find distance smaller than 0.01
% j - indices of 'grid_data' include red point

[BN]

Thank you. I really appreciate your kindness and lessons that I learned.

Kind Regards

[darova]

My pleasure!

[BN]

Dear darova,

Hope you are well.

I have a little problem here, I would be grateful if you can help me. The accepted answer code works perfectly with my attached data to this question.

I wanted to use this code for another data sets with different latitude and longitude values that I attached now in this comment, But I don't know why some issues happen during doing that.

May I ask you please look at my new attached data and guide me on how to modify the code? Because I have some problem as I listed below:

1. As this figure below shows, my first comment problem occurred here again (four green points Being involved even when I delete 0.25. I want to select only one green point that has redpoint(s) (based on 0.5 x 0.5 degree that I mentioned that exist for each green point) (All green points are one centroid (center of a pixel, when the pixel is in 0.5 x 0.5-degree resolution (square))) I want something like grey boxes.

2. To solve this problem I try to add 0.25 in these lines:

x = floor( (p(:,3)-24.25)/0.5 )*0.5+24.25; 
y = floor( (p(:,2)-44.25)/0.5 )*0.5+44.25; 

at the first look, I thought it, okay but then I saw in some area, the wrong box draw over red points for example here:

when I measurement right and left, and up and down distance of green point by ruler I saw that they are not equal (0.5 * 0.5)

Now I don't know why this code does not work well for my new data.

Do you know how should I do?

Thank you so much again

[darova]

I believe it happens because of new mesh (0.025)

>> g(1,:)
ans =
   44.0250   24.0250

Little improvement

x0 = g(1,2) + 0.25;
y0 = g(1,1) + 0.25;
x = floor( (p(:,2)-x0)/0.5 )*0.5+x0;
y = floor( (p(:,3)-y0)/0.5 )*0.5+y0;

looks equal

[BN]

Thank you so much