Counting program every positive element

Hello.

I have problem with one program and I even doesn't know how to start.

I have an array which is built with zeros and positive numbers. I want to know how many points there are between starting zero and zero ending of "the isle of positive numbers". To show my issue, I gave the image of one of my arraies and I want to get from it (from array which create this image...) matrix which have 3 columns (1col. start date/start point; 2col. end date/end point; 3col. how many points are between start date and end date N_i=...). The number of the rows is determined by how many "isles of positive numbers" have each array.

[Andrei Bobrov]

Please attach file with your data (small example of your data).

[Milosz Mart]

clear all; close all;

filename = 'C:\[...]\kwazi_data_1993.txt'

delimiter = '\t';

formatSpec = '%f%f%f%f%f%f%[^\n\r]';

fileID = fopen(filename,'r');

dataArray = textscan(fileID, formatSpec, 'Delimiter', delimiter, 'EmptyValue' ,NaN, 'ReturnOnError', false);

fclose(fileID);

data = dataArray{:, 1};

clearvars filename delimiter formatSpec fileID dataArray ans;

filename = 'C:\[...]\kwazi_data_1993.txt';

delimiter = '\t';

formatSpec = '%{yyyyMMddHH}D%*s%*s%*s%*s%*s%[^\n\r]';

fileID = fopen(filename,'r');

dataArray = textscan(fileID, formatSpec, 'Delimiter', delimiter, 'EmptyValue' ,NaN, 'ReturnOnError', false);

fclose(fileID);

DATA = dataArray{:, 1};

clearvars filename delimiter formatSpec fileID dataArray ans;

filename = 'C:\[...]\calk1993.txt'

delimiter = '\t';

formatSpec = '%f%f%f%f%f%f%[^\n\r]';

fileID = fopen(filename,'r');

dataArray = textscan(fileID, formatSpec, 'Delimiter', delimiter, 'EmptyValue' ,NaN, 'ReturnOnError', false);

fclose(fileID);

VarName1 = dataArray{:, 1};

clearvars filename delimiter formatSpec fileID dataArray ans;

for k=1:365

    if VarName1(1:k,1)>0
        KKK=VarName1(1:k,1);
    else 
        KKK=VarName1(1:k,1)>=0
    end
end
KK=KKK.*VarName1

[Guillaume]

Please remove all the blank lines you've inserted between each line of code, then select all the code and press the {}Code button. That will make your post a lot more readable.

Answer 1

[Guillaume]

This is the common problem of finding the run length of a sequence. There are plenty of answers for this on this forum.

First, identify the sequence. For you it's positive numbers,

v = [0, rand(1, 20), -rand(1, 5), rand(1, 12), 0, rand(1,2)]; %demo data.
insequence = v > 0;

Points in the sequence are indicated by 1, points not in the sequence by 0. Therefore the start of a sequence is found when 0 is followed by 1, the end when 1 is followed by 0

transitions = diff([0 insequence 0]);

transitions is 1 for start of sequence, -1 for end of sequence. I've put 0 at each end to make sure you still have [0 1] or [1 0] if the sequence starts or ends with 1.

startsequences = find(transitions == 1)
endsequences = find(transitions == -1) - 1

Because of the 0 added on each end, you're guaranteed to have the same number of startsequences and endsequences

Answer 2

[Andrei Bobrov]

Let date1 - arrey with your dates, DATA - your data

f = fopen('kwazi_data_1993.txt');
c = textscan(f,'%s','delimiter','\n');
fclose(f);
date1 = c{:};
DATA = importdata('calk1993.txt');

i1 = [(1:numel(DATA))', bwlabel(DATA > 0)];
lo = i1(:,2) > 0;
a = accumarray(i1(lo,2),i1(lo,1),[],@(x){[min(x),max(x),numel(x)]});
b = cat(1,a{:});

out = [date1(b(:,1:2)), num2cell(b(:,3))];

[Andrei Bobrov]

corrected

Answer 3

[Image Analyst]

To determine how many positive values are between index1 and index2 of your signal, you can do this

numPositives = sum(yourSignal(index1:index2) > 0);

If you have a bunch of these "runs" of positive values and don't know the starting and stopping indexes, then you can use regionprops to get the areas and indexes

labeledSignal = bwlabel(yourSignal > 0);
measurements = regionprops(labeledSignal, 'Area', 'PixelIdxValues');
allLengths = [measurements.Area];

allLengths will be a list of the lengths of each "run" of positive values.