Why is my script not finishing when I run from command line?

Hello,

I wrote a script to process a large data file and write a subset of data out to csv file. I run the file from command line as:

 >matlab -nodisplay < script.m

When I first started running it a few weeks ago it completed in a few minutes but now it hangs and does not show it has completed when in fact it has. Can anyone tell me why? I am using an older version of matlab - R2013a (8.1.0.604) 64-bit (glnxa64).

Andre

Answer 1

[Walter Roberson]

Enclose the code in your script in a try/catch block, and at the end of your script put an explicit "quit"

Historically MATLAB did not always exit when it reached end of file on standard input, especially when it encountered an error that crashed the normal flow of control.

You are a bit safer to use

matlab -nodisplay -r "try; run('YourScript.m'); end; quit"

[Andre]

Below is the code I was running. Your suggestions while they should work are not for whatever reason. I should add that I am working on a remote linux cluster and the data are mounted on a drive and I am writing to another mounted drive.

            % Andre Pattantyus 
            % May 19, 2016
            % Code to process WRF downscaling output to obtain climate data.

            cd /share/Volume03/wrfout/2057/12

            % Finding the number of files in the directory
            WRFdir = dir('wrfout*'); % Finds the needed files
            num_cycles = length(WRFdir); % Turns that information into a number
            CFHT_timeseries_file=zeros(num_cycles*24,9); %length of month in hrs for 9 vars
            CFHT_name = 'CFHT_sfcVariables_'; % Making a Text String
            %
            %~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
            %
            DperR=180/pi; % degrees per radians
            L=2.453e6; % latent heat of vaporization J/kg
            Rv=461; % gas constant for moist air J/kg
            A=2.53e8*1000; % conversion constant in Pa for Td calculation
            B=5.42e3; % conversion constant in K for Td calculation

            AA=6.116441;
            MM=7.591386;
            Tn=240.7263;

            m=0;

            n = 1; % The counter index
            while n <= length(WRFdir) % Start of the loop, will work while n<= the number of files
            name = getfield(WRFdir(n),'name'); % Get?s the name of the file
            yrmn= name(12:18);
            %~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
            % GET Variables
            CFHT_filename_string = strcat(CFHT_name,yrmn); % stringing Name and Year
            ncid=netcdf.open(name,'NOWRITE');
            Tempdata=netcdf.inqVarID(ncid,'T2');
            Udata=netcdf.inqVarID(ncid,'U10');
            Vdata=netcdf.inqVarID(ncid,'V10');
            MRdata=netcdf.inqVarID(ncid,'Q2'); % mixing ratio at 2 meters
            PSFCdata=netcdf.inqVarID(ncid,'PSFC');
            SNOWdata=netcdf.inqVarID(ncid,'SNOWNC'); % accumulated grid scale snow and ice
            RAIN1data=netcdf.inqVarID(ncid,'RAINC'); % accumulated cumulus precipitation
            RAIN2data=netcdf.inqVarID(ncid,'RAINNC'); % accumulated grid scale precipitation
            Hour=netcdf.inqVarID(ncid,'Times');
            Temp=netcdf.getVar(ncid,Tempdata); % K
            datahour=netcdf.getVar(ncid,Hour);
            U=netcdf.getVar(ncid,Udata); % m/s
            V=netcdf.getVar(ncid,Vdata); % m/s
            MR=netcdf.getVar(ncid,MRdata); % kg/kg
            PSFC=netcdf.getVar(ncid,PSFCdata); % Pa
            SNOW=netcdf.getVar(ncid,SNOWdata); % mm
            RAIN1=netcdf.getVar(ncid,RAIN1data); % mm
            RAIN2=netcdf.getVar(ncid,RAIN2data); % mm
            %~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

            % retreive Lat and Lon grids. DO THIS ONCE AND SAVE ARRAYS FOR REFERENCE
            LAT_ind=netcdf.inqVarID(ncid,'XLAT');
            LAT_full=netcdf.getVar(ncid,LAT_ind);
            LAT=LAT_full(:,:,1);
            LON_ind=netcdf.inqVarID(ncid,'XLONG');
            LON_full=netcdf.getVar(ncid,LON_ind);
            LON=LON_full(:,:,1);

            %close netcdf file
            netcdf.close(ncid);

            %calculate wind speed and direction from components
            ws=sqrt(U.^2+V.^2); % wind speed
            dir=270-atan2(V,U)*DperR; % wind direction

            %calculate RH from mixing ratio, T2 and PSFC
            Es=610.94*exp(17.625*(Temp-273.15)./(243.04+(Temp-273.15)))/100; % from  Lawrence 2005, BAMS
            E=MR.*(PSFC/100)./(0.622+MR); % vapor pressure from mixing ratio with Pa to mb correction 
            RH=E./Es*100;

            %calculate Td from Lawrence 2005, BAMS p. 226
            Td=243.04*log((E*100)/610.94)./(17.625-log(E*100/610.94))+273.15;

            %calculate total rain
            PREC=RAIN1+RAIN2;

            % get lat and lon for CFHT = 19.82525 -155.4689
            [c ind]=min(abs(LAT(1,:)-19.82525));
            CFHT_lat=ind;
            [c ind]=min(abs(LON(:,1)--155.4689));
            CFHT_lon=ind;

            CFHT_sfcTemp=squeeze(Temp(CFHT_lon,CFHT_lat,:));
               CFHT_sfcPres=squeeze(PSFC(CFHT_lon,CFHT_lat,:));
               CFHT_sfcRH=squeeze(RH(CFHT_lon,CFHT_lat,:));
               CFHT_sfcWS=squeeze(ws(CFHT_lon,CFHT_lat,:));
               CFHT_sfcDIR=squeeze(dir(CFHT_lon,CFHT_lat,:));
               CFHT_sfcTd=squeeze(Td(CFHT_lon,CFHT_lat,:));
               CFHT_PREC=squeeze(PREC(CFHT_lon,CFHT_lat,:)); % total precipiation liquid
               CFHT_SNOW=squeeze(SNOW(CFHT_lon,CFHT_lat,:));
               CFHT_sfcQ=squeeze(MR(CFHT_lon,CFHT_lat,:)); %mixing ratio

            p=size(Temp,3);

            if n==1
               for t=1:p
                  CFHT_timeseries_file(m+t,1)=CFHT_sfcTemp(t);
                  CFHT_timeseries_file(m+t,2)=CFHT_sfcPres(t);
                  CFHT_timeseries_file(m+t,3)=CFHT_sfcRH(t);
                  CFHT_timeseries_file(m+t,4)=CFHT_sfcQ(t);
                  CFHT_timeseries_file(m+t,5)=CFHT_sfcTd(t);
                  CFHT_timeseries_file(m+t,6)=CFHT_sfcWS(t);
                  CFHT_timeseries_file(m+t,7)=CFHT_sfcDIR(t);
                  CFHT_timeseries_file(m+t,8)=CFHT_PREC(t);
                  CFHT_timeseries_file(m+t,9)=CFHT_SNOW(t);
               end
            n=n+1;
            m=m+24;
            elseif n==num_cycles
               for t=1:p-1
                  CFHT_timeseries_file(m+t,1)=CFHT_sfcTemp(t);
                  CFHT_timeseries_file(m+t,2)=CFHT_sfcPres(t);
                  CFHT_timeseries_file(m+t,3)=CFHT_sfcRH(t);
                  CFHT_timeseries_file(m+t,4)=CFHT_sfcQ(t);
                  CFHT_timeseries_file(m+t,5)=CFHT_sfcTd(t);
                  CFHT_timeseries_file(m+t,6)=CFHT_sfcWS(t);
                  CFHT_timeseries_file(m+t,7)=CFHT_sfcDIR(t);
                  CFHT_timeseries_file(m+t,8)=CFHT_PREC(t);
                  CFHT_timeseries_file(m+t,9)=CFHT_SNOW(t);
               end
            else
               for t=2:p
                  CFHT_timeseries_file(m+t,1)=CFHT_sfcTemp(t);
                  CFHT_timeseries_file(m+t,2)=CFHT_sfcPres(t);
                  CFHT_timeseries_file(m+t,3)=CFHT_sfcRH(t);
                  CFHT_timeseries_file(m+t,4)=CFHT_sfcQ(t);
                  CFHT_timeseries_file(m+t,5)=CFHT_sfcTd(t);
                  CFHT_timeseries_file(m+t,6)=CFHT_sfcWS(t);
                  CFHT_timeseries_file(m+t,7)=CFHT_sfcDIR(t);
                  CFHT_timeseries_file(m+t,8)=CFHT_PREC(t);
                  CFHT_timeseries_file(m+t,9)=CFHT_SNOW(t);
               end
            n=n+1;
            m=m+24;
            end
               dlmwrite(['/share/Volume01/wrfout/',CFHT_filename_string,'.csv'],CFHT_timeseries_file,'delimiter',',','precision',15);
            end

[Walter Roberson]

I suggest writing name to a log file so you can trace how far it is getting. In most programming languages I would say "and remember to flush the buffer", but in MATLAB that is automatically done unless you give special permissions when you open the file (either 'A' or 'W' instead of 'a' or 'w')