Hi -
The scat2grid functions assume that the x and y locations of the scattered points contain no missing data. It's the data field, not the locations that are missing, after all. We'll fix the function so it'll detect any missing values among xpts, and ypts, and just skip over those in the calculation as it does for missing values in argument 3, the data values. But in the meantime you can work around this by defining your xpts and ypts without multiplying the locations by 0*swh_grid:
let xpts = xsequence(x[g=swh_grid] + 0*y[g=swh_grid])
let ypts = xsequence(x[g=swh_grid] + y[g=swh_grid])The other cases you did which worked fine must somehow not quite have the same kinds of definitions. It's not the fact that it's from a netCDF dataset, but some variation in the definitions of the variables.
let datapts = xsequence(swh_grid)
let xscale=0.1
let yscale=0.1
let cutoff=2
let swh_smooth = scat2gridgauss_xy(xpts,ypts,datapts,x[g=fine_grid],y[g=fine_grid],xscale,yscale,cutoff,0)
Ansley
On 1/28/2011 4:27 PM, Szymon Roziewski wrote:I have to use this or similar function because I get horrible result. I discussed that problem with Dr. Hein and he advised me to use scat2gridgauss_xy. Then I had cdf data and all worked properly. But now when I try to reuse it it doesn't work. The script looks like being hanged.
I give the link to what Dr. Hein said.
http://www.pmel.noaa.gov/maillists/tmap/ferret_users/fu_2010/msg00803.html
Kind regards,
Szymon Roziewski
2011/1/28 Ansley Manke <Ansley.B.Manke@xxxxxxxx>
Hi -
Is Ferret crashing, or the script exiting? Or does it just take a very long time? Why are you using the scat2grid function? It seems that you're starting with gridded data and regridding it to this very fine grid. The Ferret regridding transformations would do this directly.
Once you have read in your data, and defined the fine grid,
yes? let swh_smooth = swh_grid[gx=xfine,gy=yfine]
yes? fill swh_smooth
This regrids almost at once. There are several regridding transformations to choose from; run the Ferret command SHOW TRANSFORMS command for a list of them.
Ansley
On 1/28/2011 4:44 AM, Szymon Roziewski wrote:
Dear Ferreters,
I have such a script:
set win/size=6.0/aspect=`1.1`
SET VIEWPORT FULL
CANCEL MODE LOGO
DEFINE AXIS/X=-5.666666667:5.25:0.083333333/unit=degree xlon
DEFINE AXIS/Y=-2.0:10.0:0.083333333/unit=degree ylat
DEFINE GRID/x=xlon/y=ylat grid_xlon_xlat
FILE/VARIABLES=swh_grid/COLUMNS=132/GRID=grid_xlon_xlat swh20101018090000.txt
set var/bad=-1 swh_grid
set var/units=meters swh_grid
set var/title="Significant Wave Height" swh_grid
DEFINE AXIS/X=-5.666666667:5.25:0.010416667/unit=degree xfine
DEFINE AXIS/Y=-2.0:10.0:0.010416667/unit=degree yfine
DEFINE GRID/x=xfine/y=yfine fine_grid
let xpts = xsequence(0*swh_grid+x[g=swh_grid])
let ypts = xsequence(0*swh_grid+y[g=swh_grid])
let datapts = xsequence(swh_grid)
let xscale=0.1
let yscale=0.1
let cutoff=2
let swh_smooth = scat2gridgauss_xy(xpts,ypts,datapts,x[g=fine_grid],y[g=fine_grid],xscale,yscale,cutoff,0)
list swh_smooth
wchich should do interpolation data, but it doesn't. The script follows up to the line with list swh_smooth and as reached as is iddle.
I think the line with scat2gridgauss_xy function failed. But I don't know the reason.
Why is that I don't know. I have very similar script which copes with cdf data and do it well.
But in this case previous approach failed. (When I try using txt data with flag bad=-1.)
I attach the data file.
Kind regards,
Szymon Roziewski
--
Z wyrazami szacunku,
Szymon Roziewski
Attachment:
do2.gif
Description: GIF image
