# 4d Interpolation of Table

9 views (last 30 days)
River Vorse on 9 Nov 2020
Answered: Raunak Gupta on 13 Nov 2020
I'm trying to interpolate a set of 4 variable data values in a table. Can I use interpn() for this?
the first column ranges (0:2:10)- full table not shown since there are hundreds of values. The first three columns are inputs such that at x =0, y= 16, z=40, there are a number of outputs shown in the right 7 columns. I'm trying to interpolate so that if I have a set of inputs x = 0.5, y = 16.7, z = 37.2, I can interpolate a value for one of the outputs in the right 7 columns.

River Vorse on 9 Nov 2020
My thought is :
x = data(:,1);
y = data(:,2);
z = data(:,3);
v = data(;,4);
xp = 0.5;
yp = 16.7;
zp = 37.2;
vp = interpn(x,y,z,v,xp,yp,zp)
River Vorse on 9 Nov 2020
this gives the following error Raunak Gupta on 13 Nov 2020
Hi,
The error when using interpn is due to the data not being present in Grid format, meaning for every triplet of (x,y,z) in some range of x,y,z should have a value v also present. You can create the Grid format data for x,y,z using
[X1, X2, X3] = ndgrid(x,y,z);
But for the same you need to reshape the 4th column such that it represents corresponding values in X1, X2, X3. For this the length of unique x,y,z is required which gives the indices where values for learning interpolation are present.
You can see here about how the input array should look like.