How to use interpn on non-monotonic data

Question

Below is some part of my data. The 1st column R is the value I got from the experiment, and X Y Z are the coordinates. I am trying to use matlab n-D interpolation function. Matlab said my coordinates are not monotonic increased value. But I cant change or rearrange my coordinates.

Did I use the wrong function?

Please tell me what should I do.

R              X     Y    Z
5.05256e-18  0.016 0.015 0.032
4.99958e-18  0.016 0.015 0.064
5.04485e-18  0.016 0.015 0.128
5.49613e-18  0.016 0.0195 0.032
5.45348e-18  0.016 0.0195 0.064
5.43161e-18  0.016 0.0195 0.128
5.9393e-18  0.016 0.03 0.032
5.98785e-18  0.016 0.03 0.064
6.01929e-18  0.016 0.03 0.128
6.54936e-18  0.016 0.06 0.032
6.45947e-18  0.016 0.06 0.064
6.52379e-18  0.016 0.06 0.128
5.06516e-18  0.032 0.015 0.032
5.04897e-18  0.032 0.015 0.064
5.14022e-18  0.032 0.015 0.128
5.45993e-18  0.032 0.0195 0.032
5.4616e-18  0.032 0.0195 0.064
5.38434e-18  0.032 0.0195 0.128
5.94097e-18  0.032 0.03 0.032
5.91148e-18  0.032 0.03 0.064
6.05671e-18  0.032 0.03 0.128
6.56989e-18  0.032 0.06 0.032
6.58173e-18  0.032 0.06 0.064
6.45971e-18  0.032 0.06 0.128

This is my test code.

I want to get the interpolation value at X=0.06 Y=0.07 Z=0.08

interpn(X,Y,Z,R,0.06,0.07,0.08,'linear')

matlab message: Error using griddedInterpolant The grid vectors are not strictly monotonic increasing.

it will be helpful if you could also post the code you were using. — bla

McMa McMa · Accepted Answer · 2013-07-15T09:38:32

I am afraid that your problem goes a bit beyond the monotony of the vectors. To use interpn in this case you would also need R to be a 3 dimensional matrix. Also the values for 0.06 and 0.07 seem to be outside the range of X and Y respectively.

If you can fix the range problem, try using the function

yi = griddatan(x,y,xi)

In your case the thing would look like:

Grid=[X,Y,Z];
Point=[0.03,0.05,0.08] % with values inside the grid and maybe transposed like  Point=[0.03;0.05;0.08]; I don't know, didn't try it myself.

Result=griddatan(Grid,R,Point);

Hope that works, as I wrote above, I didn't try running it, so you might have to transpose the things here and there to get it working, but I'm almost sure that griddatan() should do the trick.

How to use interpn on non-monotonic data

1 Answers