Differences

This shows you the differences between two versions of the page.

--- internal:administration:idl [2026/03/31 16:22] – [matrix multiplication] jan
+++ internal:administration:idl [2026/04/28 08:26] (current) – [interpolation] jan
@@ Line 609: / Line 609: @@
   * One can use this to generate matrices with a variable increasing only in one dimension and arrays of function values:
-  * Alternatively one may use rebin
-  x = fltarr(Nx)
-  y = fltarr(Ny)
-  xx = rebin( x, Nx, Ny )
-  yy = rebin( y, Ny, Nx )
   ; generate Nx*Ny Matrix xx with Nx values increasing in x-direction from xa to xb
@@ Line 631: / Line 626: @@
   * As matrix operations are faster than for-loops this may speed up your code a lot ...
+  * Alternatively one may use **rebin** , but note that rebin uses bilinear interpolation (sample=0) or nearest neigbor (sample=1) and accordingly might be a bit slower
+  x = fltarr(Nx)
+  y = fltarr(Ny)
+  xx = rebin( x, Nx, Ny, sample=1 )
+  yy = transpose(rebin( y, Ny, Nx ))
 ==== procedure names ====
@@ Line 3848: / Line 3851: @@
 If you want to interpolate only over certain dimensions they must be the last ones. You may use transpose to rearrange. Lets assume you have an array depending on x, y and time, i.e. z=fltarr(Nx,Ny,Nt) and you want to interpolate to certain x and y coordinates you can do this as follows:
-  z_ipol = transpose( interpolate( transpose(z_org,[2,0,1]) , ix_ipol , iy_ipol , /grid  ), [1,2,0] )
+  z_ipol = transpose( interpolate( transpose(z_org,[2,0,1]) , ix_ipol , iy_ipol , /grid, MISSING=!VALUES.f_nan  ), [1,2,0] )
+The /GRID flag allow to use different sizes for ix_ipol and iy_ipol and the MISSING keyword sets values of z_ipol outside the ranges of z_org to the missing value (no extrapolation).