Project

General

Profile

Forums » Support »

jScanPositively wrongly set to 1 by operator samplegrid

Added by Roger Leriche about 2 years ago

In the following example, resampled.grib2 is obtained from original.grib2 by the command :
cdo samplegrid,1 original.grib2 resampled.grib2
As the factor is equal to 1, all data are kept.
When examining the files with the grib_dump utility (from EcCodes tools), we can see that the data are identical, in the same order, but jScanPositively is equal to 1 in resampled.grib2 instead of 0 in original.grib2.
Is there a reason or is it a bug ?

Download all files
original.grib2 (227 Bytes) original.grib2
resampled.grib2 (227 Bytes) resampled.grib2

Replies (10)

RE: jScanPositively wrongly set to 1 by operator samplegrid - Added by Uwe Schulzweida about 2 years ago

Hi Roger,
I can't reproduce this problem, which CDO version your are using (cdo -V)?
Cheers,
Uwe

RE: jScanPositively wrongly set to 1 by operator samplegrid - Added by Roger Leriche about 2 years ago

Hi Uwe,

I use CDO 2.0.5 in Cygwin.
I have the same behavior with CDO 1.9.9 on Ubuntu.
Below two extracts of the result of grib_dump on resampled.grib2

And the same ones on original.grib2

As you can see, jScansPositively has changed, but the latitude of first and last points have not changed. Neither the data.

Cheers,

Roger

Download all files
clipboard-202211171105-wz3kq.png (10.7 KB) clipboard-202211171105-wz3kq.png
clipboard-202211171108-oeuib.png (10.4 KB) clipboard-202211171108-oeuib.png
clipboard-202211171111-1imsy.png (8.01 KB) clipboard-202211171111-1imsy.png
clipboard-202211171112-0h7jl.png (8.09 KB) clipboard-202211171112-0h7jl.png

RE: jScanPositively wrongly set to 1 by operator samplegrid - Added by Uwe Schulzweida about 2 years ago

I couldn't reproduce the problem even with CDO version 1.9.9 and 2.0.3. It could also be due to the ecCodes version used. The version number of ecCodes can also be found in the output of cdo -V.

RE: jScanPositively wrongly set to 1 by operator samplegrid - Added by Roger Leriche about 2 years ago

EcCodes version is 2.26.0

The package libeccodes-tools (grib_dump, ...) was not installed with the cdo package for Cygwin.
I installed it later, same version

Download all files
clipboard-202211171202-pzs91.png (6.87 KB) clipboard-202211171202-pzs91.png
clipboard-202211171204-titpm.png (2.33 KB) clipboard-202211171204-titpm.png

RE: jScanPositively wrongly set to 1 by operator samplegrid - Added by Uwe Schulzweida about 2 years ago

It seems that your CDO installation was configured with --enable-hirlam-extensions. With this extension I can reproduced this problem now. This bug will be fixed in the next CDO release 2.1.1.
Cheers,
Uwe

RE: jScanPositively wrongly set to 1 by operator samplegrid - Added by Roger Leriche about 2 years ago

Hi Uwe,

For information, today I was creating a GRIB test file with the "input" operator (cdo -f grb input,r15x15 out.grb < my_ascii_data).
In the out.grb file, the jScansPositively was set to 1 too.
I guess it is the same problem.

Cheers,

Roger

RE: jScanPositively wrongly set to 1 by operator samplegrid - Added by Uwe Schulzweida almost 2 years ago

Hi Roger,

The latitutes of the predefined CDO grid r15x15 have the range -90 to +90 degrees. So I think jScansPositively=1 is correct in this case.

Cheers,
Uwe

RE: jScanPositively wrongly set to 1 by operator samplegrid - Added by Roger Leriche almost 2 years ago

Hi Uwe,

You are right ! In this case, jScansPositively=1 is correct.
But jScansPositively is also set to 1 with the attached grid definition.

Cheers,

Roger

grid_15_15.txt (294 Bytes) grid_15_15.txt

RE: jScanPositively wrongly set to 1 by operator samplegrid - Added by Uwe Schulzweida almost 2 years ago

Hi Roger,

yes, this is the same problem.

Cheers,
Uwe

    (1-10/10)