Khadija please look at this video

https://www.youtube.com/watch?v=mnoGmS2XtDg

@Kiprotich thanks I have already watched it.

I am still working on my code to resolve my issue.. it actually generate what I want but when I tried to visualize it with Panoply it only gave me arrays with no plotting, but when I used qgis I had the layer and the information aside

Hii Khadija Kaimous

In python, please check the command fillna()

First, open the netCDF data using xarray then use fillna command

eg:

import xarray

ds=xr.open_dataset('test.nc')

ds=ds.fillna(-999)

here 999 will replace the NaN values

Thanks Mr Aravind

But I am working with netCDF4, so is there a way to do it in this module ?

Hii Khadija Kaimous

The structure of both netcd4 and xarray are the same. I think xarray has some advantages over netcdf4.

I don't know any module available in netcdf4 for your need

Hii Khadija Kaimous

Please check this programe

import xarray as xr

import netcdf4

ds=xr.open_dataset('test.nc')

ds1=ds.fillna(-999)

ds2=ds1.to_netcdf('test_new.nc')

In test_new.nc, NaN values are stored as -999, so after you can open this NetCDF in netcdf4

This may work for your need

Hi thanks for all your answers,

I need to calculate the average temperature of in each position (x,y) over the 12 bands and then store them in the same position in a netCDF file and if there is an invalid data or NaN data it should be appeared in the output file as well.

I succeeded to generate the output I need but I only worked in a peace of the netCDF I didn't include all X,Y data because of the time execute that lasts more than an hour. But I usually use Panoply to verify and visualize the result.. I got another issue that I only find the array without plotting so I don't know if if my code really does what I want or not

list1=[]

list2=[]

result=0

j=0

k=0

i=0

for y in range(all_data.shape[2]):

for x in range(all_data.shape[1]):

for t in range(all_data.shape[0]):

if str(data1_var[t,y,x])!='-9999.0' and str(data1_var[t,y,x])!='nan':

i+=1

result+=float(str(data1_var[t,y,x]))

elif str(data1_var[t,y,x])=='-9999.0':

j+=1

elif str(data1_var[t,y,x])=='nan':

k+=1

if j==12:

list1+=[-9999.0]

j=0

elif k==12:

list1+=[None]

k=0

elif i!=0:

list1+=[result/i]

i=0

result=0

list2+=[list1]

list1=[]

Hii Khadija Kaimous

Would you please show me the data structure?

I would like to know how the data is arranged.

Please use ncdump -h data.nc

here is the data

Hii Khadija Kaimous

This data showing wrong values

the highest value of temperature is 500 degree celsius.

could you please verify this data

Hello Aravind

I did not actually find the highest value as 500 but it was 19.1 i used panoply to visualize it. there are some invalid data like -9999.0 and NaN values

Hi Khadija Kaimous ,

I woud'nt recommend writing such Python routines if you just want the output as another netcdf with the same attributes itself. There's another tool called the climate data operators (CDO) to do that, and its faster too :

To install cdo in Ubuntu :

sudo apt-get install cdo

Just browse to the folder in your Terminal and type the following :

To calculate the monthly mean :

cdo monmean in.nc out.nc

To calculate the yearly mean :

cdo yearmean in.nc out.nc

You can do a lot more actually, just refer the Documentation in the link below,

https://code.mpimet.mpg.de/projects/cdo/embedded/cdo.pdf

Thank you Mohammed Cassim Sannan i'll take this solution into consideration at least to help me understand more netCDF structure an then i'll code it in python

Thank you all dear reserchers for your feedbacks