NGP

       Interpolate an irregularly sampled field using Nearest Grid Point

       This function interpolates irregularly gridded points to a
       regular grid using Nearest Grid Point.

       Mathematical functions, Interpolation

       Result = NGP, VALUE, POSX, NX[, POSY, NY, POSZ, NZ,
                     /AVERAGE, /WRAPAROUND, /NO_MESSAGE]

       VALUE: Array of sample weights (field values). For e.g. a
              temperature field this would be the temperature and the
              keyword AVERAGE should be set. For e.g. a density field
              this could be either the particle mass (AVERAGE should
              not be set) or the density (AVERAGE should be set).
       POSX:  Array of X coordinates of field samples, unit indices: [0,NX>.
       NX:    Desired number of grid points in X-direction.

      POSY: Array of Y coordinates of field samples, unit indices: [0,NY>.
      NY:   Desired number of grid points in Y-direction.
      POSZ: Array of Z coordinates of field samples, unit indices: [0,NZ>.
      NZ:   Desired number of grid points in Z-direction.

       AVERAGE:    Set this keyword if the nodes contain field samples
                   (e.g. a temperature field). The value at each grid
                   point will then be the average of all the samples
                   allocated to it. If this keyword is not set, the
                   value at each grid point will be the sum of all the
                   nodes allocated to it (e.g. for a density field from
                   a distribution of particles). (D=0).
       WRAPAROUND: Set this keyword if the data is periodic and if you
                   want the first grid point to contain samples of both
                   sides of the volume (see below). (D=0).
       NO_MESSAGE: Suppress informational messages.
 Example of default NGP allocation: n0=4, *=gridpoint.
     0   1   2   3     Index of gridpoints
     *   *   *   *     Grid points
   |---|---|---|---|   Range allocated to gridpoints ([0.0,1.0> --> 0, etc.)
   0   1   2   3   4   posx
 Example of NGP allocation for WRAPAROUND: n0=4, *=gridpoint.
   0   1   2   3         Index of gridpoints
   *   *   *   *         Grid points
 |---|---|---|---|--     Range allocated to gridpoints ([0.5,1.5> --> 1, etc.)
   0   1   2   3   4=0   posx

       Prints that a NGP interpolation is being performed of x
       samples to y grid points, unless NO_MESSAGE is set.

       All input arrays must have the same dimensions.
       Postition coordinates should be in `index units' of the
       desired grid: POSX=[0,NX>, etc.

       Nearest grid point is determined for each sample.
       Samples are allocated to nearest grid points.
       Grid point values are computed (sum or average of samples).

       nx = 20
       ny = 10
       posx = randomu(s,1000)
       posy = randomu(s,1000)
       value = posx^2+posy^2
       field = ngp(value,posx*nx,nx,posy*ny,ny,/average)
       surface,field,/lego

       Use tsc.pro or cic.pro for a higher order interpolation schemes.    A
       standard reference for these interpolation methods is:   R.W. Hockney
       and J.W. Eastwood, Computer Simulations Using Particles (New York:
       McGraw-Hill, 1981).

       Written by Joop Schaye, Feb 1999.
       Check for LONG overflow  P. Riley/W. Landsman   December 1999

[Home Page] [Software, Documentation] [IDL Documentation] [Quick Reference] [Feedback]