Variable not printing to csv

Hello,

I have a script which cycles through a datafile called 'MRS_struct' (produced from an analysis software). Then it gets certain variables from the file and exports the values to csv. I need to link these values to an identifier-- in the script called 'PID'. However, my variable 'PID' will not print to the csv. Is it too long? I tried truncating the PID_cellarray but it did not work. PID is taking the field called 'metabfile' within MRS_struct which is pretty long (~150 characters).

Also, to add, when I index 'PID_cellarray' or 'PID' I get the output I want, I just can't seem to get it to export to csv.

Here is the code:

% This script goes through the fit folder, finds files with 'MRS_struct'
% then goes into those .m files and goes down the layers to find the
% data we want-- can add more if needed.

%All are based off of MRS_struct file from QUANTIFY
%Go down layers from here: MRS_struct.out.vox1....

clc;
clear;
clearvars;
%addpath (pwd);

%%Paths and setup-- need to change paths for each batch extraction
%%%%change these based on study/ project file locations%%%%
source_path = 'B:\BBA\Gannet\Tests\Test_batchAllSteps\Test_batchWithWater_7.26.18\GannetQuantify_output'; %where we are pulling data (.mat files) from
final_path = 'B:\BBA\Gannet\Tests\Test_batchAllSteps\Test_batchWithWater_7.26.18\GannetQuantify_output'; %where to output excel file

keyword = 'MRS_struct';

%need to declare variables/arrays as zero so we can add to them in loops
cd(source_path)

%%Make empty arrays with N_max, 1 dimension to hold N data.
%1, N_max is row; N_max,1 is column

%Setup + PID
N_max = 100;
count = 0;
PID_cellarray = cell(N_max, 1);
MAT_filename_cellarray = cell(N_max,1);

%Empty cell arrays here that we will fill later
%Data input in comment to right of empty array line below

%GABA cell arrays
GABACr_cellarray  = cell(N_max,1); %from output: ConcCr
GABAIU_cellarray = cell(N_max,1);%from output: ConcIU
GABAFitError_cellarray = cell(N_max,1);%from output: FitError
GABA_CSFCorr_cellarray = cell(N_max, 1);%from output: ConcIU_CSFCorr
GABA_TissCorr_cellarray = cell(N_max, 1);%from output: ConcIU_TissCorr
GABA_AlphaTissCorr_cellarray = cell(N_max, 1); %from output: ConcIU_AlphaTissCorr
GABA_AlphaTissCorr_GrpNorm_cellarray = cell(N_max, 1); %from output: ConcIU_AlphaTissCorr_GrpNorm

%Glx cell arrays
GlxCr_cellarray  = cell(N_max,1); %from output: ConcCr
GlxIU_cellarray = cell(N_max,1); %from output: ConcIU
GlxFitError_cellarray = cell(N_max,1); %from output: FitError
Glx_CSFCorr_cellarray = cell(N_max, 1); %from output: ConcIU_CSFCorr
Glx_TissCorr_cellarray = cell(N_max, 1); %from output: ConcIU_TissCorr
Glx_AlphaTissCorr_cellarray = cell(N_max, 1); %from output: ConcIU_AlphaTissCorr
Glx_AlphaTissCorr_GrpNorm_cellarray = cell(N_max, 1); %from output: ConcIU_AlphaTissCorr_GrpNorm

%Tissue composition cell arrays
GMFra_cellarray = cell(N_max, 1); %from output: GMfra
WMFra_cellarray = cell(N_max, 1); %from output: WMfra
CSFfra_cellarray = cell(N_max, 1); %from output: CSFfra

%output_xls_filename = 'Gannet_MAT_Output';

%%Read Data + set variables to load into arrays / export at end
display ('Step 1: look for MAT files in Quantify folder');
MAT_file_location = dir([source_path, '/*.mat']);%list all conent from location of MAT files
tmp0 = load(MAT_file_location(1).name); %loading MRS_struct file
tmp_metabfile = tmp0.MRS_struct.metabfile; %'.metabfile' layer from MRS_struct

for i = 1: length(tmp_metabfile); %will loop for # of metabfile entries in MRS_struct. I.e. number of datasets.

      if(strfind(MAT_file_location(1).name, keyword)); %if file has MRS_struct in name
       %Using (1) rather than (i) because there will only be 1 .mat file in
       %the folder. (i) confuses the indexing

          %load mat files to array- loading file we found at top of loop with keyword 
          tmp = load(MAT_file_location(1).name); % getting the name and ext. of file "MRS_struct.mat"
                                                 % tmp is now MRS_struct file
          count = count + 1;

          display ('Look for data from MRS_struct');
          PID = tmp.MRS_struct.metabfile(i); % trying to link identifier to data for output
          %dont put these values in '' or it will be string, this gives data

          display ('Getting GABA data from MRS_struct');
          %GABA Data from MRS_struct-- here we are getting values from MRS
          %struct and assigning to variables
          GABACr = tmp.MRS_struct.out.vox1.GABA.ConcCr(i);
          GABAIU = tmp.MRS_struct.out.vox1.GABA.ConcIU(i);
          GABAFitError = tmp.MRS_struct.out.vox1.GABA.FitError(i);
          GABA_CSFcorr = tmp.MRS_struct.out.vox1.GABA.ConcIU_CSFcorr(i);
          GABA_TissCorr = tmp.MRS_struct.out.vox1.GABA.ConcIU_TissCorr(i);
          GABA_AlphaTissCorr = tmp.MRS_struct.out.vox1.GABA.ConcIU_AlphaTissCorr(i);
          GABA_AlphaTissCorr_GrpNorm = tmp.MRS_struct.out.vox1.GABA.ConcIU_AlphaTissCorr_GrpNorm(i);

          %%%%here with (i) we get the first data point, without we get
          %%%%both-- but how do we extract both, or do we use (i) and cycle
          %%%%another iteration within MRS_struct...

          display ('Getting Glx data from MRS_struct');
          %Glx Data from MRS_struct
          GlxCr = tmp.MRS_struct.out.vox1.Glx.ConcCr(i);
          GlxIU = tmp.MRS_struct.out.vox1.Glx.ConcIU(i);
          GlxFitError = tmp.MRS_struct.out.vox1.Glx.FitError(i);
          GlxIU_CSFcorr = tmp.MRS_struct.out.vox1.Glx.ConcIU_CSFcorr(i);
          Glx_TissCorr = tmp.MRS_struct.out.vox1.Glx.ConcIU_TissCorr(i);
          Glx_AlphaTissCorr = tmp.MRS_struct.out.vox1.Glx.ConcIU_AlphaTissCorr(i);
          Glx_AlphaTissCorr_GrpNorm = tmp.MRS_struct.out.vox1.Glx.ConcIU_AlphaTissCorr_GrpNorm(i);

          display ('Getting Tissue data from MRS_struct');
          %Tissue Concentration Data from MRS_struct
          GMFra = tmp.MRS_struct.out.vox1.tissue.GMfra(i);
          WMFra = tmp.MRS_struct.out.vox1.tissue.WMfra(i);
          CSFfra = tmp.MRS_struct.out.vox1.tissue.CSFfra(i);

          %Here we are assigning above data to empty arrays created earlier
          %We will then export them below

          display ('Assigning data to arrays');
          %Loading GABA variables/data into arrays
          MAT_filename_cellarray{count} = MAT_file_location(1).name;
          PID_cellarray{count} = PID;
          GABACr_cellarray{count}  = GABACr;
          GABAIU_cellarray{count} = GABAIU;          
          GABAFitError_cellarray{count} = GABAFitError;
          GABA_CSFCorr_cellarray{count} = GABA_CSFcorr;
          GABA_TissCorr_cellarray{count} = GABA_TissCorr;
          GABA_AlphaTissCorr_cellarray{count} = GABA_AlphaTissCorr;
          GABA_AlphaTissCorr_GrpNorm_cellarray{count} = GABA_AlphaTissCorr_GrpNorm;

          %Loading Glx variables/data into arrays
          GlxCr_cellarray{count} = GlxCr;
          GlxIU_cellarray{count} = GlxIU;
          GlxFitError_cellarray{count} = GlxFitError;
          GlxIU_cellarray {count} = GlxIU;
          GlxFitError_cellarray{count} = GlxFitError;
          Glx_CSFCorr_cellarray{count} = GlxIU_CSFcorr;
          Glx_TissCorr_cellarray{count} = Glx_TissCorr;
          Glx_AlphaTissCorr_cellarray{count} = Glx_AlphaTissCorr;
          Glx_AlphaTissCorr_GrpNorm_cellarray{count} = Glx_AlphaTissCorr_GrpNorm;

          display ('tissue conc');
          %Loading Tissue Conc variables/data into arrays
          GMFra_cellarray{count} = GMFra;
          WMFra_cellarray{count} = WMFra;
          CSFfra_cellarray{count} = CSFfra;

    end
end

%array of headers, values-- MAKE SURE THESE ARE IN THE SAME ORDER
%put these 2 into an outputcell and write to excel

display ('Getting ready to output')
output_header_cellarray = ... %Headers for output document
[{'PID'},{'GABAConcCr'}, {'GABAConcIU'}, {'GABAFitError'}, {'GABA_CSFCorr'}, {'GABA_TissCorr'}, {'GABA_AlphaTissCorr'} , {'GABA_AlphaTissCorr_GrpNorm'} , {'GlxConcCr'} , {'GlxConcIU'} , {'GlxFitError'} , {'Glx_CSFCorr'} , {'Glx_TissCorr'} , {'Glx_AlphaTissCorr'} , {'Glx_AlphaTissCorr_GrpNorm'} ];

output_value_cellarray = ...%Data that goes into each column
[PID_cellarray, GABACr_cellarray, GABAIU_cellarray, GABAFitError_cellarray, GABA_CSFCorr_cellarray, GABA_TissCorr_cellarray, GABA_AlphaTissCorr_cellarray, GABA_AlphaTissCorr_GrpNorm_cellarray , GlxCr_cellarray , GlxIU_cellarray , GlxFitError_cellarray, Glx_CSFCorr_cellarray , Glx_TissCorr_cellarray , Glx_AlphaTissCorr_cellarray , Glx_AlphaTissCorr_GrpNorm_cellarray];

output_cell = [output_header_cellarray; output_value_cellarray]; %write header + value cellarrays

output_xls_filename = 'Gannet_Quantify_Extract';
xlswrite(output_xls_filename, output_cell);

[dpb]

...
output_xls_filename = 'Gannet_Quantify_Extract';
xlswrite(output_xls_filename, output_cell);

From

 doc xlswrite
 ...
 filename
 ...
...If you do not specify an extension, xlswrite uses the default, .xls.
 ...

So, you're not writing a .csv file but an .xls file instead.

Also, NB; you can name a file written with xlswrite something other than a recognized Excel extension (*.xls, .xlsx, etc) but naming the file 'something.csv' does not cause xlswrite to write a csv file' it still creates the same .xls format just now Excel will no longer automagically recognize the content thru the name extension.

>> r={2,pi};                          % just some simple data
>> xlswrite('xlsfile.csv')            % try to write .csv
>> [~,~,raw]=xlsread('xlsfile.csv');  % read it back
raw(1,1:2) =
  {[ 2]}    {[3.1416]}
>>

AFAIK, there is no high-level i/o routine to write a .csv file with mixed data content; csvwrite works on numeric arrays only.

[Holly Pothier]

The file format is not important, but thanks for pointing that out.

So there is no way to print a string array to column A, then numeric arrays to the rest of the Columns?

[dpb]

Depends on what file format you do want... xlswrite will write an Excel file of a cell array if each cell contains one element; the content of the cells is mixable.

writetable will write the content of a table to a file where the variables in the table can be of whatever type; the kicker is you have to start with a table.

You can write anything including cell arrays with low-level i/o calls to fprintf but one has to do the formatting explicitly and, of course, dereference the cell content correctly to get the underlying data contained in the cell array to pass to fprintf.

It is a glaring weakness in my opinion that there really isn't any higher-level manner to write cell arrays; there are auxiliary tools for reading text files of varying formatting but there's nothing of any similar nature for output.

The sort of thing one has to do is outlined for simple example (they're always simple in the doc, aren't they???) at write-to-delimited-data-files

Answer 1

[OCDER]

Found an old code of mine. Does this work?

% C = {'a', 'b', 'c'; 'a1', [2 3 4], 3};
% writeCell2CSV(C, 'tester.csv');

function writeCell2CSV(C, FileName)
C(cellfun('isempty', C)) = {''};
NumLoc = cellfun('isclass', C, 'double');
StrLoc = ~NumLoc;
C(NumLoc) = cellfun(@mat2str, C(NumLoc), 'un', 0); 
C(StrLoc) = strrep(C(StrLoc), ',', ';');
[FID, MSG] = fopen(FileName, 'w');
assert(FID > 0, '%s: Could not create/write to file "%s".\n  %s', mfilename, FileName, MSG);
Fmt = [repmat('%s,', 1, size(C, 2)-1) '%s\n'];
for j = 1:size(C, 1)
    fprintf(FID, Fmt, C{j, :});
end
fclose(FID);

[dpb]

At least a start...TMW should be watching.

[Holly Pothier]

Thank you both for the help with this. I was able to work around my problem by truncating the indexed datapoints from the PID_cellarray. I am now able to print a string variable in one of the columns.

[Holly Pothier]

Actually, truncating was not the issue. I was trying to print PID_cellarray, which was an array of the variable PID-- which I believe was indexed incorrectly. Now I am printing the variable PID_truncated, which is PID{1}(1:10), where PID = tmp.MRS_struct.mask.vox{1}.T1image(i).... I think I had my indexing off.

Thanks for the help!