[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[netCDF #YAR-472810]: NetCDF row- vs column-major I/O convention?

Hi Evan,

> This problem is relatively simple but has been driving me nuts. I'm
> using the NetCDF library in a C/C++ program to read and write data. The
> NetCDF C User's Guide says NetCDF uses C row-major convention. I am
> getting the opposite results. Here is an example...
> 
> I created a small test file that looks like this:
> 
> % ncdump an_demo_test.cdf
> netcdf an_demo_test {
> dimensions:
> Lat = 6 ;
> Lon = 12 ;
> Alt = 1 ;
> variables:
> double PredictedReflectivity(Alt, Lat, Lon) ;
> PredictedReflectivity:Units = "dBZ" ;
> data:
> 
> PredictedReflectivity =
> 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11,
> 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23,
> 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35,
> 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47,
> 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59,
> 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71 ;
> }

Yes, that shows the data in row-major order, with
the "Lon" index varying most rapidly.

> Now, I read this file into my C++ program using (where variables are
> defined accordingly):
> 
> status = nc_open("an_demo_test.cdf", NC_NOWRITE, &nc_id);
> status = nc_inq_varid(nc_id, "PredictedReflectivity", &zh_id);
> status = nc_get_var_double(nc_id, rh_id, zh_arr);
> 
> Then I see what I've got...
> 
> for(int i=0;i<Lat;i++)
> {
> for(int j=0;j<Lon;j++)
> {
> std::cout << "img[" << i << "," << j << "] = " <<
> img[i+(Lat*j)] << "\n";

It looks like you have your indices reversed in the above
expression, I think you should be using:

  img[j+(Lat*i)]

because j is varying most rapidly (the row index).

--Russ

> }
> std::cout << "\n";
> }
> 
> kamet.engr.colostate.edu % an_demo
> Dimension (Lat) read completed (ret = 0, Lat = 6)...
> Dimension (Lon) read completed (ret = 0, Lon = 12)...
> Variable declared: 6x12
> sizeX() = 6
> sizeY() = 12
> img[0,0] = 0
> img[0,1] = 6
> img[0,2] = 12
> img[0,3] = 18
> img[0,4] = 24
> img[0,5] = 30
> img[0,6] = 36
> img[0,7] = 42
> img[0,8] = 48
> img[0,9] = 54
> img[0,10] = 60
> img[0,11] = 66
> 
> img[1,0] = 1
> img[1,1] = 7
> img[1,2] = 13
> img[1,3] = 19
> img[1,4] = 25
> img[1,5] = 31
> img[1,6] = 37
> img[1,7] = 43
> img[1,8] = 49
> img[1,9] = 55
> img[1,10] = 61
> img[1,11] = 67
> 
> img[2,0] = 2
> img[2,1] = 8
> img[2,2] = 14
> img[2,3] = 20
> img[2,4] = 26
> img[2,5] = 32
> img[2,6] = 38
> img[2,7] = 44
> img[2,8] = 50
> img[2,9] = 56
> img[2,10] = 62
> img[2,11] = 68
> 
> img[3,0] = 3
> img[3,1] = 9
> img[3,2] = 15
> img[3,3] = 21
> img[3,4] = 27
> img[3,5] = 33
> img[3,6] = 39
> img[3,7] = 45
> img[3,8] = 51
> img[3,9] = 57
> img[3,10] = 63
> img[3,11] = 69
> 
> img[4,0] = 4
> img[4,1] = 10
> img[4,2] = 16
> img[4,3] = 22
> img[4,4] = 28
> img[4,5] = 34
> img[4,6] = 40
> img[4,7] = 46
> img[4,8] = 52
> img[4,9] = 58
> img[4,10] = 64
> img[4,11] = 70
> 
> img[5,0] = 5
> img[5,1] = 11
> img[5,2] = 17
> img[5,3] = 23
> img[5,4] = 29
> img[5,5] = 35
> img[5,6] = 41
> img[5,7] = 47
> img[5,8] = 53
> img[5,9] = 59
> img[5,10] = 65
> img[5,11] = 71
> 
> But if I switch my loop to...
> 
> for(int i=0;i<Lon;i++)
> {
> for(int j=0;j<Lat;j++)
> {
> std::cout << "img[" << j << "," << i << "] = " <<
> img[j+Lon*i) << "\n";
> }
> std::cout << "\n";
> }
> 
> I get the desired result...
> 
> kamet.engr.colostate.edu % an_demo
> Dimension (Lat) read completed (ret = 0, Lat = 6)...
> Dimension (Lon) read completed (ret = 0, Lon = 12)...
> Variable declared: 6x12
> sizeX() = 6
> sizeY() = 12
> ret = 0
> img[0,0] = 0
> img[1,0] = 1
> img[2,0] = 2
> img[3,0] = 3
> img[4,0] = 4
> img[5,0] = 5
> 
> img[0,1] = 6
> img[1,1] = 7
> img[2,1] = 8
> img[3,1] = 9
> img[4,1] = 10
> img[5,1] = 11
> 
> img[0,2] = 12
> img[1,2] = 13
> img[2,2] = 14
> img[3,2] = 15
> img[4,2] = 16
> img[5,2] = 17
> 
> img[0,3] = 18
> img[1,3] = 19
> img[2,3] = 20
> img[3,3] = 21
> img[4,3] = 22
> img[5,3] = 23
> 
> img[0,4] = 24
> img[1,4] = 25
> img[2,4] = 26
> img[3,4] = 27
> img[4,4] = 28
> img[5,4] = 29
> 
> img[0,5] = 30
> img[1,5] = 31
> img[2,5] = 32
> img[3,5] = 33
> img[4,5] = 34
> img[5,5] = 35
> 
> img[0,6] = 36
> img[1,6] = 37
> img[2,6] = 38
> img[3,6] = 39
> img[4,6] = 40
> img[5,6] = 41
> 
> img[0,7] = 42
> img[1,7] = 43
> img[2,7] = 44
> img[3,7] = 45
> img[4,7] = 46
> img[5,7] = 47
> 
> img[0,8] = 48
> img[1,8] = 49
> img[2,8] = 50
> img[3,8] = 51
> img[4,8] = 52
> img[5,8] = 53
> 
> img[0,9] = 54
> img[1,9] = 55
> img[2,9] = 56
> img[3,9] = 57
> img[4,9] = 58
> img[5,9] = 59
> 
> img[0,10] = 60
> img[1,10] = 61
> img[2,10] = 62
> img[3,10] = 63
> img[4,10] = 64
> img[5,10] = 65
> 
> img[0,11] = 66
> img[1,11] = 67
> img[2,11] = 68
> img[3,11] = 69
> img[4,11] = 70
> img[5,11] = 71
> 
> It looks like the NetCDF file is being read into a column-major
> variable. Can you help me figure this out???
> 
> Many thanks,
> Evan
> 
> 

Russ Rew                                         UCAR Unidata Program
address@hidden                      http://www.unidata.ucar.edu



Ticket Details
===================
Ticket ID: YAR-472810
Department: Support netCDF
Priority: High
Status: Closed