Merging Multiple text files

I have multiple text files, each file is data for 1 day with 3 columns, 1 columns for longtitude, 1 columns for latitude and 1 columns for precip value.

It looks like that

 108.650    13.950     0.000
 108.450    13.383     0.000
 105.833    22.150     0.000
 106.217    21.300     -99.000
 104.283    22.533     0.000
 105.717     9.283     0.100

The value of longtitude and latitude is same for all file.

Now, I want to merge all file to 1 netcdf or text file with dimension is 455x455x365 (455 is the number of long and lat, 365 is precip value for each day of a year)

How can I do that?

[Geoff Hayes]

minh - do you have 365 files (one for each day of the year)? Does each file have 455 rows? Please clarify.

[minh lan]

Geoff Hayes,

Yes, I have 365 files and each file has 455 rows and 3 columns,

Answer 1

[KSSV]

txtfiles = dir('*.txt') ;
ncfile = 'myfile.nc' ;
N = length(txtfiles) ;
nx = 455 ; ny = 455 ;    % number of lat, lon
for i = 1:N
    data = importdata(txtfiles(i).name) ;
    lon = data(:,1) ;
    lat = data(:,2) ;
    p = data(:,3) ;
    if i == 1
        % Do interpolation
        idx = boundary(lon,lat) ;
        xi = linspace(min(lon),max(lon),nx) ;
        yi = linspace(min(lat),max(lat),ny) ;
        [X,Y] = meshgrid(xi,yi) ;
        idx = inpolygon(X,Y,lon(idx),lat(idx)) ;
        Z = griddata(lon,lat,p,X,Y) ;
        Z(~idx) = NaN ;
        % Longitude
        nccreate(ncfile,'lon','Dimensions',{'lon',1,nx}) ;
        ncwrite(ncfile,'lon',xi) ;
        ncwriteatt(ncfile,'lon','long_name','longitude');
        % Latitude
        nccreate(ncfile,'lat','Dimensions',{'lat', 1, ny})
        ncwrite(ncfile,'lat',yi)
        ncwriteatt(ncfile,'lat','long_name','latitude');
        % Precipitation
        nccreate(ncfile,'time','Dimensions',{'time',1,Inf}) ;
        nccreate(ncfile,'p','Dimensions',{'lon',nx,'lat',ny,'time',Inf})
        ncwriteatt(ncfile,'p','long_name','precipitation');
        
        ncwrite(ncfile,'time',i,i) ;
        ncwrite(ncfile,'p',Z,[1,1,i])
    end
    Z = griddata(lon,lat,p,X,Y) ;
    Z(~idx) = NaN ;
    
    ncwrite(ncfile,'time',i,i) ;
    ncwrite(ncfile,'p',Z,[1,1,i])
end

[minh lan]

KSSV,

p = reshape(data(:,3),nx,ny) ;

why you need reshape here? And when I run code, It has error.

"To RESHAPE the number of elements must not change."

And I see nx, ny is the number and data(:,3) is matrix 455x1

[KSSV]

As you said, you have 455 lats and lons.....I though your data is structured. Attach one file.

[minh lan]

I attach one file of a day below. And I have 365 file like this

[KSSV]

YOu data is a scatter data. YOu want to save it as scaatered data ot grid?

[minh lan]

I want to save it as grid data.

[KSSV]

Edited the code.

[minh lan]

I still have error:

'lon' exists in file. Can not modify existing variables.

How can I fix it?

[KSSV]

YOu have to delete the previous nc file...if it exists.

[minh lan]

Thank you so much.

[minh lan]

One more question,

If I want to have scatter data to compare the different map between the scattered data and the data after interpolation?

How can I change your code?

Thank you.

[KSSV]

USe lon,lat,p for scattered data.....use X,Y,Z for grid data.