alex-robinson / ncio Goto Github PK

I was trying to write a plain integer (one value) into a netCDF file. Using nc_write. According to the API:

filename        name of the NetCDF file in which to write data
name            name of the variable in NetCDF file to be written
dat             Fortran data type of data to be written
dims            vector of dimension names of the variable in NetCDF file (optional)
dim1,...,dim6   individual dimension names of the variable in NetCDF file (optional)
start           vector of values specifying starting indices for reading data from each dimension (optional)
count           vector of values specifying how many values to read in each dimension (optional)
long_name       NetCDF attribute, a long descriptive name of variable (optional)
standard_name   NetCDF attribute specifying the CF convention standard name of the variable (optional)
grid_mapping    name of the grid this variable is mapped on (optional)
units           NetCDF attribute of the units of the variable (optional)
missing_value   Value of missing data to be written to file (optional)
ncid            File ID for a file that remains open for various NCIO calls (optional) 

... the only thing to provide are the first three arguments.

Thus using the simplest test possible:

program test
   use ncio

   call nc_create("test.nc",overwrite=.true.,netcdf4=.true.)
   call nc_write("test.nc","int_var",1)

end program test

But this example compiles fine but do not run, error message is:

ncio:: nc_create   :: test.nc
 ncio:: error: NetCDF: Invalid dimension ID or name
STOP stopped by ncio.

Within the module it should be caught by nc_write_int_pt that calls nc4_write_internal, but I could not find the case where ndims=0 in the latter. I guess it defaults to ndims=1 and then crashes since none exist/specified.

The netCDF4 library allows one or more 'unlimited' dimensions so that arrays can be appended along them. Could the authors further implement the 'append' mode and allows the usages of multiple 'unlimited' dimensions?

When I try to write data into a non-existent file, the error message is misleading:

ncio:: error: NetCDF: String match to name in use

Example:

call nc_create     ("test.nc")
call nc_write_dim  ("nothere.nc", 'x', x=0, dx=1, nx=10)

ncio:: error: NetCDF: Variable not found

It would be nice to know, which variable causes this error.
I have looked into the source and it appears that this error message comes from the underlying netcdff library, so it should be fixed there, but maybe ncio could catch this error and provide additional information until then?

When using the debugging option I have a crash in nc_put_att because of an un-allocated variable (see below). This only happens when I've exported the environment variable debug = 1.

This doesn't seem to be a big problem, but it makes using ncio in debug mode a pain or even impossible.

I'm on a mac (10.15.7) using gfortran from macports (gcc version 11.2.0 (MacPorts gcc11 11.2.0_1)).

Monaco:ncio einola$ export debug=1
Monaco:ncio einola$ make test
gfortran -w -g -p -ggdb -ffpe-trap=invalid,zero,overflow,underflow -fbacktrace -fcheck=all -I.obj -J.obj -I/opt/local/include -c -o .obj/ncio.o ncio.f90
gfortran -w -g -p -ggdb -ffpe-trap=invalid,zero,overflow,underflow -fbacktrace -fcheck=all -I.obj -J.obj -I/opt/local/include -c -o .obj/ncio_transpose.o ncio_transpose.f90
gfortran -w -g -p -ggdb -ffpe-trap=invalid,zero,overflow,underflow -fbacktrace -fcheck=all -I.obj -J.obj -I/opt/local/include -o test_ncio.x .obj/ncio.o .obj/ncio_transpose.o test_ncio.f90 -L/opt/local/lib -lnetcdff -lnetcdf
 
    test_ncio.x is ready.
 
Monaco:ncio einola$ ./test_ncio.x

 ====== WRITING ======

ncio:: nc_create   :: out_ncio.nc
ncio:: nc_write_map:: out_ncio.nc : polar_stereographic
At line 822 of file ncio.f90
Fortran runtime error: Allocatable argument 'v' is not allocated

Error termination. Backtrace:
#0  0x102fe76e0 in ???
#1  0x102fe7f78 in ???
#2  0x102fe84eb in ???
#3  0x102e1e46b in nc_put_att
        at /Users/einola/tmp/ncio/ncio.f90:822
#4  0x102e154cf in nc_write_dim_internal
        at /Users/einola/tmp/ncio/ncio.f90:1833
#5  0x102e17b48 in __ncio_MOD_nc_write_dim_int_pt
        at /Users/einola/tmp/ncio/ncio.f90:1605
#6  0x102e3401e in test
        at /Users/einola/tmp/ncio/test_ncio.f90:74
#7  0x102e3b447 in main
        at /Users/einola/tmp/ncio/test_ncio.f90:5

PS: Thanks for the code - it's a lifesaver!

This project is a very useful tools. But when I read a variable of a Group, I write like "nc_ndims(filename,"/G1/Longitude")", program will raise a error "ncio:: error: NetCDF: Variable not found".So what should I do to read variables of Group?
thanks

The ncio module was written in a single file to make it portable and easy of use, consistent with the overall idea of ncio. However, as pointed out by a reviewer, this advantage may become an obstacle for the developer when the source code grows in size, especially because of the necessary repetition of the same functional code for various variable types making up a single fortran interface.

A first and simple step, as suggested by the reviewer, could be separating the various functionality into separate files, and automatically aggregate the pieces into the user-ready ncio.f90 (for example as a Makefile command, possibly even a git hook on commit).

One may go further in this direction and reduce the amount of human-maintainable source code to type-independent prototype code, when possible, for instance by using the capability of the python language for processing strings (e.g. format method), or by using existing tools. Similarly, a Makefile command would process and aggregate the "source" code into a single ncio.f90 with full interfaces.

Thanks for making this. I hate the verbosity of the netcdf interface, so had always meant to write something like this. It is a welcome relief that someone else has beaten me to it.

I'm interested in creating an object-oriented interface, with a user defined netcdf type that can be automatically populated and queried.

Is the best approach to make a separate module (say ncio_oo) that uses the module as-is?

I would like to see more map projections added to NCIO. Specifically lambert conformal conic and gnomonic grids