Ask matlab to compare variables per row between two column? Output a number "1" if the values are the same, output number "0" if the values are different.

dataFolder = uigetdir();%select folder
filePattern = fullfile(dataFolder, '*.csv');
list = dir(filePattern);

Output = table();
for kk = 1:numel(list)
    filename = list(kk).name;
    data = readtable(fullfile(list(kk).folder, filename));
    if numel(list)(:,1)==numel(list)(:,4)
     disp('Correct')   
     value = 1;
    elseif  numel(list)(:,1)~=numel(list)(:,4)   
     disp('Incorrect')   
     value = 0;
    end
end

For multiple .csv files in a folder that appear like this:

I want to ask matlab to compare variables per row between column 1 and 4.

Output a number "1" if the values are the same:

numel(list)(:,1)==numel(list)(:,4)

and "0" if they are not:

numel(list)(:,1)~=numel(list)(:,9)

For example row 2 column 1 = 6, column 4 = NaN.

It means I would like to output a column in wich in the row 2 I will have a 0.

Where am I wrong? Something with the syntax for sure but where?

[awezmm]

If you compare two NaN's, do you want it to be equal and return 1 or non-equal and return 0?

[Myke Ziz]

I would like to be equal to 1 yes.

Only when not the same equal to 0, this for numbers and words.

Answer 1

[Walter Roberson]

(Output{:,1} == Output{:,4}) | (isnan(Output{:,1}) & isnan(Output{:,4}))

The second part is needed to implement the "nans are equal" criteria.

[Myke Ziz]

Thank you for yor answer as always very kind of yours in helping!

Should I not switch 'Output' with 'numel(list)' in order for the code to work?

dataFolder = uigetdir();%select folder
filePattern = fullfile(dataFolder, '*.csv');
list = dir(filePattern);

Output = table();
for kk = 1:numel(list)
    filename = list(kk).name;
    data = readtable(fullfile(list(kk).folder, filename));
    if (Output{:,1} == Output{:,4}) | (isnan(Output{:,1}) & isnan(Output{:,4}))
     disp('Correct')   
     value = 1;
    elseif  (Output{:,1} ~= Output{:,4}) | (isnan(Output{:,1}) & isnan(Output{:,4}))
     disp('Incorrect')   
     value = 0;
    end
end

Thank you in advance!

M.Z.

[Walter Roberson]

No.

You asked to compare per row . The result of the expression I gave will be a per-row vector, not a scalar. You would not use if with it, because if is only true in the case that all of the entries are non-zero.

mask = (data{:,1} == data{:,4}) | (isnan(data{:,1}) & isnan(data{:,4}));
value = all(mask);
some = any(mask);
if value
    fprint('All Correct for file "%s"\n', filename);
elseif some
    fprintf('Some Correct for file "%s"\n', filename);
else
    fprintf('None Correct for file "%s"\n', filename);
end
frac = mean(mask);

but is that even what you want??

What do you even want to store in Output ? Perhaps

Output.value(k) = value;
Output.Accuracy(k) = 100*frac;
Output.filename{k} = filename;

[Myke Ziz]

I try to compare all the elements between 2 columns.

Comparaison will occur by same row elements between the 2 columns.

As output, 1 will be assigned if the elements were identical, 0 if they were not.

I would like to create 1 single output table for all the .csv files I have in the folder (1 column per .csv file comparaison of the 2 columns).

Sorry for my writing I am not a nativ english speaker.

[Walter Roberson]

value(kk) = all((data{:,1} == data{:,4}) | (isnan(data{:,1}) & isnan(data{:,4})));

Afterwards,

Output = array2table(value, 'VariableNames', {list.name});

[Myke Ziz]

I am trying to insert value(kk) in the code but still choosing the wrong place I guess.

Could you show me how would you write the full code whenever you have time.

Thank you a lot Robert!

dataFolder = uigetdir();%select folder
filePattern = fullfile(dataFolder, '*.csv');
list = dir(filePattern);

Output = table();
for kk = 1:numel(list)
    filename = list(kk).name;
    data = readtable(fullfile(list(kk).folder, filename));
    if ........   
     value = 1;
    elseif  ........   
     value = 0;
    end
end

value(kk) = all((data{:,1} == data{:,4}) | (isnan(data{:,1}) & isnan(data{:,4})));

Output = array2table(value, 'VariableNames', {list.name});

[Walter Roberson]

dataFolder = uigetdir();%select folder
filePattern = fullfile(dataFolder, '*.csv');
list = dir(filePattern);
numfiles = length(list);
value = zeros(numfiles,1);
for kk = 1:numfiles
    filename = list(kk).name;
    data = readtable(fullfile(list(kk).folder, filename));
    value(kk) = all((data{:,1} == data{:,4}) | (isnan(data{:,1}) & isnan(data{:,4})));
end

If you have R2019a or earlier,

[~, basenames] = cellfun(@fileparts, {list.name});
varnames = matlab.lang.makeValidName(basenames);
varnames = matlab.lang.makeUniqueStrings(varnames, {}, namelengthmax);

If you have R2019b or later,

varnames = {list.name};

Then either way,

Output = array2table(value, 'VariableNames', varnames);

[Myke Ziz]

I am trying like this but it is still wrong.

dataFolder = uigetdir();%select folder
filePattern = fullfile(dataFolder, '*.csv');
list = dir(filePattern);
numfiles = length(list);
value = zeros(numfiles,1);

for kk = 1:numfiles
    filename = list(kk).name;
    data = readtable(fullfile(list(kk).folder, filename));
    value(kk) = all((isequal(data{:,1},data{:,9})) | (cellfun(@isnan,(data{:,1}),'UniformOutput',false) & cellfun(@isnan,data{:,9},'UniformOutput',false)));
end
%cellfun(@isnan,a)
[~, basenames] = cellfun(@fileparts, {list.name});%1
varnames = matlab.lang.makeValidName(basenames);%2
varnames = matlab.lang.makeUniqueStrings(varnames, {}, namelengthmax);%3

%varnames = {list.name}; %at the place of 1, 2 and 3 if 2019b version or later
Output = array2table(value, 'VariableNames', varnames);

ERROR:

Undefined operator '&' for input arguments of type 'cell'.

[Walter Roberson]

Why did you switch to cellfun and why did switch to column 9? The data extract you posted does not include column 9 so I do not know what it looks like.

[Myke Ziz]

I get an error for '==' and 'isnan'.

I was trying to fix those errors.

(I switched column 4 with the 9, the more important to me would be the number 9)

With your code right now in the workspace, 'data' for column 1 and column 9 are the following ones:

[Walter Roberson]

Character vectors are never nan and so do not need to be tested for nan. The nan test is only because nan==nan is false in MATLAB rather than true

For character vector content you probably need to change the == in the first part of the test to strcmp()

[Myke Ziz]

Using this code it works, but there is something wrong in the loop, I don't obtain the full

output table.

dataFolder = uigetdir();%select folder
filePattern = fullfile(dataFolder, '*.csv');
list = dir(filePattern);
numfiles = length(list);
value = zeros(numfiles,1);

for kk = 1:numfiles
    filename = list(kk).name;
    data = readtable(fullfile(list(kk).folder, filename));
    s1 = (data{:,1});
    s2 = (data{:,9});
    tf = strcmp(s1,s2)
    %value(kk) = (strcmp(s1,s2));
end

[Walter Roberson]

value(kk) = all(tf) ;

[Myke Ziz]

With the finished code I get an error and the output table of all the columns doesn't pop up:

Code:

dataFolder = uigetdir();%select folder
filePattern = fullfile(dataFolder, '*.csv');
list = dir(filePattern);
numfiles = length(list);
value = zeros(numfiles,1);

for kk = 1:numfiles
    filename = list(kk).name;
    data = readtable(fullfile(list(kk).folder, filename));
    s1 = (data{:,1});
    s2 = (data{:,9});
    tf = strcmp(s1,s2)
    value(kk) = all(tf)
end

[~, basenames] = cellfun(@fileparts, {list.name});
varnames = matlab.lang.makeValidName(basenames);
varnames = matlab.lang.makeUniqueStrings(varnames, {}, namelengthmax);

Output = array2table(value, 'VariableNames', varnames);

Error:

Error using cellfun
Non-scalar in Uniform output, at index 1, output 1.
Set 'UniformOutput' to false.

[Walter Roberson]

[~, basenames] = cellfun(@fileparts, {list.name}, 'uniform', 0);

[Myke Ziz]

Thank you Walter!

I receive two errors for those lines:

varnames = matlab.lang.makeValidName(basenames);
varnames = matlab.lang.makeUniqueStrings(varnames, {}, namelengthmax);

Output = array2table(value, 'VariableNames', varnames);

Error using table.init (line 401)
The VariableNames property must contain one name for each variable in the table.

Error in array2table (line 64)
    t = table.init(vars,nrows,rownames,nvars,varnames);

[Walter Roberson]

value = zeros(1,numfiles);

Though it might be necessary to take special care because of the vector being passed in. I will need to reread the appropriate documentation.

[Walter Roberson]

https://www.mathworks.com/matlabcentral/answers/498717-extract-the-displayed-column-vectors-into-table-and-csv-file#comment_782904