Find values and replace them with NaN, add total number of NaN values.

Hello! So within 3 different columns of data which I have defined as:

   mintemp = b(:,4)
   maxtemp = b(:,5)
   avgtemp = b(:,6)

I want to take each individual row (1 column at a time) and find the -9999 values which are NaN values and replace them with 'NaN' so that when I calculate the average of one it doesn't skew the actual value, or find a way to calculate the average only using positive integers in Matlab if there is this function. Additionally, I would like to be able to keep a count of the total number of -9999 (NaN) values in that column so that I know how many missing values I have in the calculated average.

n=length(source_files);

for j= 1:n
mintemp = b(:,4)
NaN = -9999
ff= find(mintemp==NaN);
%This is about as far as I've got so far
f=ff+1

Any suggestions are appreciated! Thanks for your help everyone!

Answer 1

[Dan Seal]

To replace all -9999 values with NaN, you can do:

mintemp(mintemp == -9999) = NaN;

If you have Statistics Toolbox, you can then use NANMEAN:

nanmean(mintemp)

If you don't have Statistics Toolbox, you can take the mean of the values that are not NaN:

mean(mintemp(~isnan(mintemp)))

If you don't want to replace things with NaN, and want to compute the average for only positive numbers, you can do this all in one command:

mean(mintemp(mintemp>0))

[Dan Seal]

Oh, and to get a count of the -9999 values use:

sum(mintemp == -9999)

Or if you first converted to NaN, count the NaNs using:

sum(isnan(mintemp))

[Theodore]

Awesome! All of those work perfectly for calculating the mean! Thanks a ton Dan!

[Dan Seal]

Feel free to accept this answer if it solved your problem :)

[Theodore]

Sorry had to refresh to see the comment on getting a count for my -9999/NaN values, which is the only other thing I was waiting on. Worked like a charm! Thanks a ton again!