Why would I be getting "NaN" from a matrix I entered?

Hey all,

I seem to be getting a "NaN" from a matrix I entered as a variable while using the function griddata. Is there any obvious reason for this? Possibly difficulty interpolating a value?

[Patrik Ek]

You have nan in your matrix? Your matrix was created with nan. Your input data is wrong? You are a victim of a practical joke involving team mates and corruption of data? Your computer decided that it wanted a retirement?...

Sorry, but no one is all knowing. The does most likely come from a bug in the implementation. Then we need to know where it went wrong. There is many ways to get nan, so the problem is what you did to get nan.

Answer 1

[Star Strider]

We need to see your code. My guess is that you’re asking it to extrapolate.

[Jack B]

Thanks for the reply. I believe you are right. I looked at my matrix I entered and the values I have are slightly outside the range of my matrix. Thanks for the pointer.

Can I extrapolate those values afterwards?

[Star Strider]

My pleasure!

The griddata function won’t allow you to extrapolate, but interpn and some of its friends will. (The interpolation-extrapolation functions have different capabilities.) You may have to experiment to find a function that does what you want. Fortunately, they seem to share the same general argument list syntax.

There are several helpful links to other functions at the end of the griddata and other documentation pages that can help guide your search.

[Sean de Wolski]

scatteredInterpolant supports extrapolation with a few different methods.

Answer 2

[Image Analyst]

Maybe you entered the letter "o" instead of zero?