Finding and removing NaN values in a matrix

Hi Guys,

I am dealing with matrices of the size 1268*100. I wish to eliminate certain rows from this matrix. I do certain operations to get integer values for the rows that I want to keep and NaN value for the ones I dont. How can I get the rows that have NaN values and use these numbers to eliminate the rows that I dont want in my main matrix.

Eg. if my column matrix has the following values= [1 2 3 NaN 3 4 NaN 5 7 NaN]

So I will need to eliminate rows 4 7 and 10 in my main matrix.

How do I go about doing that.

Thanks, Nancy

Answer 1

[Walter Roberson]

Is it only necessary to refer to column 1, or are there other columns that might have nans that signal removal ?

If column 1 alone is enough, then you can use

values(isnan(values(:,1)),:) = [];

[Pavan]

Works well

[Shubham Maurya]

try

    values(find(isnan(values)))=[];

[Walter Roberson]

Shubham Maurya, that is testing for nan in all of the columns an removing individual nan values rather than entire rows. The result would be a column vector because you cannot leave "holes" in a rectangular matrix.

[Ahmed]

Try this

f(isnan(f))=[];

[Walter Roberson]

Ahmed, that has the same problems I described for Shubham Maurya's suggestion.