Finding and replacing numbers with NaN

Hello,

I have a matrix where the numbers only made up of 9 are fill values and should be replaced by NaN.;

For eg: There are numbers in all sorts of combinations : 99.999, 99999.99, 9999.9 99.999999, 999999.99 ....

A bit lost on how to proceed,

Making a list like below is not easy because there are n number of combinations which is not known

P(P==9999999.99)=NaN;

P(P==999999.99)=NaN;

P(P==9999.99)=NaN;

P(P==99999.99)=NaN;

Answer 1

[Image Analyst]

Try ismembertol(). Something like (untested)

Lia = ismembertol(P, 9, 0.0001);
P(Lia) = NaN;
Lia = ismembertol(P, 9.9, 0.0001);
P(Lia) = NaN;
Lia = ismembertol(P, 9.99, 0.0001);
P(Lia) = NaN;

Hopefully there is just a limited number of cases, like 10 or so. If you really might have dozens of possibilities then you should create a list somehow, like perhaps a for loop with sprintf() and str2double

for k1 = 1 : 10
	str = repmat('9', [1, k1])
	for k = 1 : length(str)
		str2 = [str(1:k), '.', str(k+1:end)]
		num = str2double(str2)
	end
end

[SChow]

Many thanks, making a list like you suggested works wonders

Answer 2

[Cris LaPierre]

I would suggest using the standardizemissing function. I don't think you are going to be able to get around having to make a list, though. I can't think of a good way for MATLAB to recognize numbers just containing 9's.

[SChow]

Thanks for your answer,

I came across standardizemissing, however as you say making a list of all possible numbers made up of all 9s is very difficult.,