Finding the first element of a matrix that satisfy a condition in a frequently manner

Hello,

I have a large matrix of data (10000x6). As you can see in the following I am imposing a condition on my data, and this condition is true in a frequent manner. The thing is I want to extract just the first element whenever the condition is true (the marked points in the image) and exclude the other elements which satisfy the condition. I would be very thankful if someone could tell me how can I do this.

Thank you,

Shahin

y_coordinate=D(6:N_atom:end,:); TR=zeros(size(y_coordinate)); aa=1; for jj=1:length(y_coordinate) if y_coordinate(jj,2)<31.35 && y_coordinate(jj,4)<0 TR(aa,:)=y_coordinate(jj,:); aa=aa+1; end end

[Adam Danz]

There is no image.

Answer 1

[Adam Danz]

... I want to extract just the first element whenever the condition is true...

This line below will return the index value of the first row that is true.

% Find the first index where condition is true (empty if none are true)
firstTrueIdx = find(y_coordinate(:,2)<31.35 & y_coordinate(:,4)<0, 1);

Now you can use firstTrueIdx| to identify the row number of that element. If it's empty, that means there were no true values.

Answer 2

[Bruno Luong]

Your loop is equivalent to

condmeet = y_coordinate(:,2)<31.35 & y_coordinate(:,4)<0;
TR = y_coordinate(condmeet ,:);

If you want to find the first element then

ifirst = find(condmeet,1,'first');
FirstTR = y_coordinate(ifirst ,:);

it returns empty array if no row meets the condition.

[Adam Danz]

condmeet and ifirst can be combined in one line which would then be the same as the other answer here.

[Bruno Luong]

Not sure about your comment Adam: the other answers need loop (with index jj), mine does not. Or you make a type with jj?

[Adam Danz]

ah, yeah... didn't mean to include the jj. good catch.

Answer 3

[shahin mohammadnejad]

Dear Adam,

Thank you for your response. I think because the plot of my data was not attached in my question, you couldn't catch my problem properly. Using the commands that you suggested, I get only the first element which satisfies the the condition. The thing is as you can see in the attached image, my data satisfy the condition many times (two of them are shown in the image), and I want whenever the condition is true I capture the first true value. For example, in this image I want the indices of both indicated points. Thank you in advance for your time and help.

Regards,

Shahin

[shahin mohammadnejad]

Dear Bruno,

My previous comment was referring to the commands suggested by you. I mentioned Adam's name at the beginning of my comment by mistake.

[Adam Danz]

"...I want whenever the condition is true I capture the first true value. For example, in this image I want the indices of both indicated points."

@Shahin, this doesn't make sense to me. Either you want the first true value or you want all true values. Which one is it?

The 2nd input to find() is the number of non-zeros to return. If this value is 1 as in both answers, it will return the first result. If you want to find all 'true's, don't use the 2nd input. Without the 2nd input, the index of all true results will be returned.

[shahin mohammadnejad]

@Adam, as an explanation of the system that I am working on: here I am tracking a gas particle's trajectory in a system, and I want to know its different features, exactly at the timestep which it passes a certain distance towards a surface. Since this particle crosses many times the initiall indicated distance, therefore, I need to capture the first true value in a frequentlly manner. I hope that now it makes sense for you what I am looking for to do here.

However, thanks to Bruno's and your previous comments I defined a loop which gives me what I want as follow:

condmeet= y_coordinate(:,2)<31.35 & y_coordinate(:,4)<0;

TR = zeros(size(y_coordinate));

for j=1:length(condmeet)-1

    if condmeet(j+1)-condmeet(j)==1

        TR(j+1,:)=y_coordinate(j+1,:);

    end

end

Here, TR is a matrix with zeros except in the rows which particle initially crosses the specified distance.

[Adam Danz]

Nice work. You can, however, avoid the loop entirely if you want. You just have to find the index values of all true values in condmeet, then differentiate the index values to detect when there are neighboring indices, then remove the earlier index of the neighboring pairs.

condmeetIdx = find(condmeet); 
deltaIdx = [diff(condmeetIdx),0];
condmeetIdx(deltaIdx==1) = [];
TR = false(size(y_coordinate));
TR(condmeetIdx) = true;

[shahin mohammadnejad]

Yes, you are right. But by using a loop, I can directly substitute the features of particles from y_coordinate matrix to TR matrix.

[Adam Danz]

Oh, that could be achieved by changing the last line of my code above.

condmeetIdx = find(condmeet); 
deltaIdx = [diff(condmeetIdx),0];
condmeetIdx(deltaIdx==1) = [];
TR = zeros(size(y_coordinate));
TR(condmeetIdx,:) = y_coordinate(condmeetIdx, :);