How do I eliminate index error in sum function.

I have imported a csv file into matlab and want to take the sum of every 48 lines. I have the following lines. When I sum the values I get an error "Index exceeds matrix dimensions". I tried increasing the size of the matrix which alleviated the problem, but I no longer was able to get the correct sum. I also have not define sum or nan sum as a variable in previous lines

clear i j nansum sum
if for i = 1:48:length(dtkp)
 h2o_daily(j) = nansum(h2o_mm(i:i+47)); %refer to something outside of matrix?
 co2_daily(j) = nansum(co2_mgm2(i:i+47));
 h2o_daily_mn(j) = nanmean(h2o_mm(i:i+47)); %mean ignoring nan values
 co2_daily_mn(j) = nanmean(co2_mgm2(i:i+47));
 dt_daily(j) = min(dtkp(i:i+1));
 tallydaily_nanh2o(j) = sum(isnan(h2o_mm(i:i+47)));%count number of nan vales
 tallydaily_nanco2(j) = sum(isnan(co2_mgm2(i:i+47)));   
 j = j+1;
end

Answer 1

[dpb]

clear i j nansum sum
if for i = 1:48:length(dtkp)
  h2o_daily(j) = nansum(h2o_...

You've cleared j but reference it before defining it...

j=0;
if for i = 1:48:length(dtkp)
  j=j+1;
  h2o_daily(j) = nansum(h2o_...

If I understand your wish, using Matlab internal storage order allows for such sums or other functions to be computed more easily--

nAvg=48;   % number of which to average, etc., ...
nVar=2;    % number of variables (columns) in data array
data=[h2o_mm(:) co2_mgm2(:)];  % put data into one array by column
data=reshape(data,nAvg,[]);    % break at 48-line intervals
stat=nanmean(data);            % compute by column (over nAvg values)
stat=reshape(stat,length(stat)/nVar),[]);  % reshape back to the nVar columns

Adjust for the other desired statistics...

[bikefan]

Thanks, this solved the issue.

However, in the data files - I do not always have a number of rows divisible by 48. What would I add to take each window of 48 and the last window be a remainder?

[dpb]

Use modulo arithmetic on the computation of number of rows...

fix(length(data)/nAvg)

is the number of full segments;

mod(length(data),nAvg)

is size of the remainder.

How to treat is your call...average over that size, ignore, ...

[bikefan]

Thanks again - How do I set the average of each parameter as a variable similar to the previous code?

[dpb]

I recommend against making a proliferation of variables in Matlab; build an array that holds the statistics and simply reference into it.

Or, if you're adamant you must have a name, look at the dataset or structures with dynamic names. This is much more programming-friendly than keep a bunch of separate variables.

You've already seen one advantage of the use of arrays--you can get the same computation for all of them "free" for the cost of one. This is a feature not to be underestimated in its importance as you begin with Matlab.