Issue with using loop

I have 3 dimensional matrix, 135*129*43099. 43099 being the daily rainfall value for 118 years(1901 to 2018). I intend to convert this daily rainfall value to monthly rainfall value, resulting matrix 135*129*1416.

For this I created a matrix of size 1416*5. Each row representing each month for 118 years. First column is year, second column is month, third column is number of number of days in a month (irrevelant and not used here), fourth and fifth column is the index. Gilmpse of such matrix's small portion is below.

Name of matrix: 'statarrayD'

1901 1 31 1 31

1901 2 28 32 59

1901 3 31 60 90

1901 4 30 91 120

1901 5 31 121 151

1901 6 30 152 181

Here is the deal, I want to add 1st to 31st value of those 43099 values to place it into a first row, similarly for second row, i want to add 32nd to 59th value of 43099 values. This way i would have 1416 rows in the resulting matrix, formed as a result of adding those 43099 values, which is named MRain here.

Code:

MRain=NaN(135,129,1416);
x=1;
for i=1:1416
    MRT=NaN(135,129,43099);
    min=statarrayD(i,4);
    max=statarrayD(i,5);
    MRT(:,:,min:max)=temp(:,:,min:max);
    MRain(:,:,x)=sum(MRT,3);
    x=x+1;
end

MRT is the temporary matrix that would return monthly values. Say when value of i=1 then, it would return the 1st to 31st rainfall values, remaining being NaN. min and max would return 1 and 31 respectively. min and max are basically 4th and 5th column of the matrix shown above. I am using sum function to add the MRain montly values, to get one value for each month.

This code seems to work perfectly without loop, when i use the numerical value like 1, 2, 3 instead of value of i, like code shown below. But returns NaN value for entire MRain matrix with used with loop. I am missing something in the loop.

%test%
MRest=NaN(135,129,1416);
MRTest=NaN(135,129,35);
min=statarrayD(1,4);
max=statarrayD(1,5);
MRTest=temp(:,:,min:max);
MRest(:,:,1)=sum(MRTest,3);

[SALAH ALRABEEI]

Try without the MRTest

MRain(:,:,x) =sum(temp(:,:,min:max),3);

[Suman Dhamala]

Wow!!! worked like wonders. Thankyou so much.

[dpb]

Look into tables, timetables, datetime, grouping variables with findgroups and splitapply

There are tools built for these kinds of analyses that, with a datetime variable associated with thd data can make these things a cakewalk...

[Jan]

Hint: avoid using "min" and "max" as variables, because shadowing built-in functions cause unexpected behavior frequently, if you try to use the functions later:

clear all
max(1:10)  % 10
max = 17;
max(1:10)  % error