Skip to content

Instantly share code, notes, and snippets.

@guziy
Last active June 30, 2022 04:59
Show Gist options
  • Star 9 You must be signed in to star a gist
  • Fork 1 You must be signed in to fork a gist
  • Save guziy/8543562 to your computer and use it in GitHub Desktop.
Save guziy/8543562 to your computer and use it in GitHub Desktop.
A quick way to copy variables and metadata from one netcdf file to another using netcdf4-python
# -*- coding: utf-8 -*-
from netCDF4 import Dataset
#input file
dsin = Dataset("crop.nc")
#output file
dsout = Dataset("crop.nc3", "w", format="NETCDF3_CLASSIC")
#Copy dimensions
for dname, the_dim in dsin.dimensions.iteritems():
print dname, len(the_dim)
dsout.createDimension(dname, len(the_dim) if not the_dim.isunlimited() else None)
# Copy variables
for v_name, varin in dsin.variables.iteritems():
outVar = dsout.createVariable(v_name, varin.datatype, varin.dimensions)
print varin.datatype
# Copy variable attributes
outVar.setncatts({k: varin.getncattr(k) for k in varin.ncattrs()})
outVar[:] = varin[:]
# close the output file
dsout.close()
@ksaha79
Copy link

ksaha79 commented Jul 17, 2020

I need my file in netCDF4 classic. It seems .setncatts won't work with netCDF4 Classic, is there a work around.

@guziy
Copy link
Author

guziy commented Jul 18, 2020

Hi @ksaha79:

I doubt the problem is with the netCDF4 library. I am guessing your python version might not support dict comprehensions.

Does this work?

outVar.setncatts({"test": "test-value"})

If yes, then you would need to replace the line with:

tmp = {}

for k in varin.ncattrs():
    tmp[k] = varin.getncattr(k)

outVar.setncatts(tmp)

@guziy
Copy link
Author

guziy commented Jul 18, 2020

Otherwise, I would need a minimal example demonstrating the problem, to be able to help you further.

Cheers

@ksaha79
Copy link

ksaha79 commented Jul 19, 2020

Adding the four lines again works fine with NETCDF3_CLASSIC, however, when I try to use the same for NETCDF4_CLASSIC format I get the following error:

lat 720
lon 1440
nv 2
time 1
lat float32
lon float32
time int32
lat_bnds float32
lon_bnds float32
analysed_sst int16
Traceback (most recent call last):
File "toto.py", line 51, in
outVar.setncatts(tmp)
File "netCDF4/_netCDF4.pyx", line 4153, in netCDF4._netCDF4.Variable.setncatts
File "netCDF4/_netCDF4.pyx", line 1617, in netCDF4._netCDF4._set_att
File "netCDF4/_netCDF4.pyx", line 1887, in netCDF4._netCDF4._ensure_nc_success
AttributeError: NetCDF: Attempt to define fill value when data already exists.

Strange is the code stops suddenly when the Fillvalue of "analysed_sst" is about to be filled and gives an error. Till then all the variable attributes are fine. Also rest of the two more variables also is not copied in the new file.

@guziy
Copy link
Author

guziy commented Jul 19, 2020

I have just tested a python3 version of the script on a file I have:

https://github.com/guziy/PyNotebooks/blob/master/netcdf/test_copy.ipynb

Is it possible for you to share the problematic file? Just a small part of it that causes the error.

@ksaha79
Copy link

ksaha79 commented Jul 19, 2020

The input files I am using can be found here: ftp://ftp.nodc.noaa.gov/pub/data.nodc/ghrsst/GDS2/L4/GLOB/NCEI/AVHRR_OI/v2/2020/085/20200325120000-NCEI-L4_GHRSST-SSTblend-AVHRR_OI-GLOB-v02.0-fv02.0.nc

@guziy
Copy link
Author

guziy commented Jul 19, 2020

I have updated the example, by adding a special treatment for the _FillValue attribute.

https://github.com/guziy/PyNotebooks/blob/master/netcdf/test_copy.ipynb

I wanted to point out that there is maybe a more elegant way of doing it now with xarray:

with xarray.open_dataset("test.nc") as ds:
       # do things with ds1 = f(ds)
       ds1.to_netcdf("copy_test.nc")

@ksaha79
Copy link

ksaha79 commented Jul 19, 2020

Thanks That worked for me.
Cheers

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment