How do I create subscript indicies to be real positive integers or logicals?

I'm attempting to get data from a .nc file and when running this code I am getting the error of using real positive integers or logicals. Each variable produces exactly what is expected except for spei producing "NaN. The values that should be returned are decimals. How do I get them to return the decimal?

clear
  clc
  loc=xlsread('Tree_Ring_Data_WI_test.xlsx');%Load treering data only file. SiteNo. species column removed
  year=loc(:,4); % formatting
  lat=loc(:,2);%latitude
  lon=loc(:,1);%longitude
  [lonf,latf]=TransLatLon(lon,lat);%Function that translate lat and lon into coordination counts

    files=dir('*.nc');%load Temperature files
    for y=1:length(files);%Number of years in file
    temp=ncread(files(y).name,'spei'); % variable name in climate file
    %ind=find(year==y+1900);
    %lont=lonf(ind);
    %latt=latf(ind);

    startm = 12 * (year - 1901)
    spei = zeros(1,1380);
    for i=1:length(lat);
         for  j=1:1382;
              spei(j)=(temp(lonf(i),latf(i), (j + startm(i)));
        end  
    end

[Star Strider]

Please use the [{}Code] button to format your code.

[Walter Roberson]

It is not clear why you initialize spei to 1380 columns, but store up to column 1382 ?

You have a nested for loop, with i outside and j inside. Every iteration of i, you overwrite all of spei, so the effect at the end is as-if you had done only the last iteration. If you have a nested for loop, you should expect to need a multidimensional output (unless you are doing something like totaling the entries.)

You use lonf(i) and latf(i) in every iteration of for j, even though they do not change over the j. It would be more efficient to store them in variables inside the for i and use the variables in the for j

Are you certain that TransLatLon can only produce positive integers? Are you certain that loc(:,4) can only contain integers that are at least 1901 ? What shows up for

nnz(isnan(temp))

?

[Ayo Deas]

That was my error to have them initializing the columns to 1380 and storing them to 1382. I adjusted that error.

The expectation of the multidimensional array is accurate. The data is stored in a multidimensional array.

Would this be a good example?

startm = 12 * (year - 1901);
    spei = zeros(1,1380);
    for i=1:length(lat)

        spei = zeros(1,1380);
        loc(startm(i),30)=spei(1,1,:); 

        for  j=1:1380
                spei(j)=(temp(lonf(i),latf(i), startm(i)));
        end

      end

I believe that TransLatLon can only produce positive integers or I could be incorrect. In this example lonf and latf both have positive integers. This is the script for TransLatLon:

function [lonf,latf]=TransLatLon(lon,lat)

lat90=90-lat;

lon180=lon+180;

latf=round(lat90/0.5)+1;

lonf=round(lon180/0.5)+1;

end

I am certain that loc(:,4) can only contain integers that are at least 1901.

nnz(isnan(temp))

ans =

     266109131

I'm unsure if it would help to know or if it makes a difference that the data from the spei file has decimals inside the cells.

[Walter Roberson]

spei(j)=(temp(lonf(i),latf(i), startm(i)));

probably needs to be

spei(i, j) = temp(lonf(i),latf(i), startm(i));

The nan in the output spei would appear to be due to the nan on the input file.

It is not uncommon for ncf files to have a 'MissingData' value, which in your case is probably NaN: the nan output would then correspond to places where there was no input.

[Ayo Deas]

Here is the updated code, taking into consideration all of the suggested actions. As far as nan on the input file, what do you mean by the "input file"? Does it make a difference that the data is stored in a .csv file as opposed to a .xls file and being placed back onto an .xls file that is within the .nc files?

From what I can see on one of the .csv files, there is data(decimals) in every cell in the first column.

Running this code has changed all the values in spei from nan to 0 but still giving the same error message of 'Subscript indices must either be real positive integers or logicals.' which I believe is progress.

 clear
  clc
  loc=xlsread('Tree_Ring_Data_WI_test.xlsx');%Load treering data only file. SiteNo. specises colum removed
  year=loc(:,4); % RWI-year change to fit your sheet formate
  lat=loc(:,2);%latitude
  lon=loc(:,1);%longitude
  [lonf,latf]=TransLatLon(lon,lat);%Function that translate lat and lon into coordination counts

files=dir('*.nc');%load Temperature files
for y=1:length(files)%Number of years in your climate file
    temp=ncread(files(y).name,'spei'); % variable name in climate file
    %ind=find(year==y+1900);
    %lont=lonf(ind);
    %latt=latf(ind);

      startm = 12 * (year - 1901);
      spei = zeros(1,1380);
      for i=1:length(lat)

        spei = zeros(1,1380);
        loc(startm(i),30)=spei(1,1,:); 
          loc(startm(i),31)=spei(1,1,:);
          loc(startm(i),32)=spei(1,1,:);
          loc(startm(i),33)=spei(1,1,:);
          loc(startm(i),34)=spei(1,1,:);
          loc(startm(i),35)=spei(1,1,:);
          loc(startm(i),36)=spei(1,1,:);
          loc(startm(i),37)=spei(1,1,:);
          loc(startm(i),38)=spei(1,1,:);
          loc(startm(i),39)=spei(1,1,:);
          loc(startm(i),40)=spei(1,1,:);
          loc(startm(i),41)=spei(1,1,:);

          for  j=1:1380
                  spei(i,j)=(temp(lonf(i),latf(i), startm(i)));
          end

        end
  end

[Walter Roberson]

I suggest that at the command line you give the command

dbstop if error

and then run the code. When it stops, examine lonf(i) and latf(i) and startm(i)