How do I get a structure to include all the data that I want?

Question

Khush Patel 2022-9-2

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1793810-how-do-i-get-a-structure-to-include-all-the-data-that-i-want

评论： Jan 2022-9-3

folder_list = dir([data_dir filesep currentfolder(3).name]);

this is my code. I want folder_list to grab everything from 3 to 7. if I use a simple for loop from 3:7, it just gives me the data for 7.

5 个评论
显示 3更早的评论隐藏 3更早的评论

Khush Patel 2022-9-2

编辑：Walter Roberson 2022-9-2

在 MATLAB Online 中打开

%%step 1 access and explore data
% Code provided.
clear
addpath(genpath(pwd));
data_dir = fullfile(pwd, 'Human_Activities', 'Selected_Data_from_HMP_Dataset');
% Code provided.
currentfolder=dir(data_dir);
%%step 2 preprocess data and extract features
folder_list = dir([data_dir filesep currentfolder(3).name]);
% Code Provided.
reference_table = table();
Label = table();
% Code Provided.
for ifolder = 1:length(folder_list)  
    % Code Provide.
    if folder_list(ifolder).bytes > 0
        fid = importdata(folder_list(ifolder).name);
        % Code Provided.
        
        %In the following code, we apply the extractAccFeatures function to 
        % generate a reference_table which stores the features from each 
        % accelerometer data in separate rows.
        reference_table = [reference_table; extractAccFeatures(fid)];
        % Code Provided.
        Label = [Label; {currentfolder(3).name}];
    end 
end
% Code Provided.
Label.Properties.VariableNames = {'Activities'};
reference_table = [reference_table, Label]

Stephen23 2022-9-3

编辑：Stephen23 2022-9-3

"I want folder_list to grab everything from 3 to 7."

That looks suspiciously like a buggy attempt to ignore the dot directory names:

https://www.mathworks.com/matlabcentral/answers/345841-unexplained-error-on-dir#answer_271627

https://www.mathworks.com/matlabcentral/answers/21150-function-call#comment_46042

https://www.mathworks.com/matlabcentral/answers/394807-create-text-file-containing-list-of-file-names#comment_556641

https://www.mathworks.com/matlabcentral/answers/288815-gui-for-loop-not-working-correctly#answer_224956

https://www.mathworks.com/matlabcentral/answers/484430-using-struct-dir-selpath-what-do-and-mean-as-struct-name#answer_395754

https://www.mathworks.com/matlabcentral/answers/40949-omitting-pointers-when-listing-folder-contents#answer_50568

https://www.mathworks.com/matlabcentral/answers/13978-list-of-files-sorting#answer_19278

As Walter Roberson wrote in that last link: "In short: if your code assumes that '.' and '..' are the first two entries in a directory, your code has a bug (even in MS Windows). If your code assumes that directory entries are returned in any sorted order, your code has a bug (in all OS.)"

Jan 2022-9-3

在 MATLAB Online 中打开

@Khush Patel: Do not do this:

addpath(genpath(pwd));

Adding the folder, which contains data files, to Matlab's PATH has no advantage, but severe disadvantages. Matlab's PATH should include only the folders containing the M-files of the used functions. Use absolute path names for data files instead:

fid = importdata(fullfile(data_dir, folder_list(ifolder).name));

I do not see in your code the command matching the description "if I use a simple for loop from 3:7".

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Walter Roberson 2022-9-2

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1793810-how-do-i-get-a-structure-to-include-all-the-data-that-i-want#answer_1040880

在 MATLAB Online 中打开

For example, to read all png files in all subfolders of data_dir, storing the input as a cell array of cell arrays broken up by folder.

dinfo = dir(data_dir);
dinfo(~[dinfo.isdir]) = [];  %remove non-folders
dinfo(ismember({dinfo.name}, {'.', '..'})) = [];  %remove . and ..
foldernames = fullfile({dinfo.folder}, {dinfo.name});
numfolders = length(foldernames);
folder_images = cell(numfolders, 1);
for D = 1 : numfolders
    thisfolder = foldernames{D};
    finfo = dir(thisfolder, '*.png');
    filenames = fullfile({finfo.folder}, {finfo.name});
    numfiles = length(filenames);
    these_images = cell(numfiles, 1);
    for F = 1 : numfiles
        thisfile = filenames{F};
        these_images{F} = imread(thisfile);
    end
    folder_images = these_images;
end

5 个评论
显示 3更早的评论隐藏 3更早的评论

Khush Patel 2022-9-2

在 MATLAB Online 中打开

Any hints to how I could apply it to this code?

%%step 1 access and explore data
% Code provided.
clear
addpath(genpath(pwd));
data_dir = fullfile(pwd, 'Human_Activities', 'Selected_Data_from_HMP_Dataset');
% Code provided.
currentfolder=dir(data_dir);
%%step 2 preprocess data and extract features
folder_list = dir([data_dir filesep currentfolder(3).name]);
% Code Provided.
reference_table = table();
Label = table();
% Code Provided.
for ifolder = 1:length(folder_list)  
    % Code Provide.
    if folder_list(ifolder).bytes > 0
        fid = importdata(folder_list(ifolder).name);
        % Code Provided.
        
        %In the following code, we apply the extractAccFeatures function to 
        % generate a reference_table which stores the features from each 
        % accelerometer data in separate rows.
        reference_table = [reference_table; extractAccFeatures(fid)];
        % Code Provided.
        Label = [Label; {currentfolder(3).name}];
    end 
end
% Code Provided.
Label.Properties.VariableNames = {'Activities'};
reference_table = [reference_table, Label]

Walter Roberson 2022-9-2

在 MATLAB Online 中打开

Adjust ThingToIgnore and OtherThingToIgnore to be the names of the objects in the directory that you want to ignore. Do not count on them being in a particular relative position.

data_dir = fullfile(pwd, 'Human_Activities', 'Selected_Data_from_HMP_Dataset');
currentfolder = dir(data_dir);
currentfolder(~[currentfolder.isdir]) = [];  %remove non-folders
currentfolder(ismember({currentfolder.name}, {'.', '..', 'ThingToIgnore', 'OtherThingToIgnore'})) = []; %remove . and .. and two things not of interest
folder_list = fullfile( {currentfolder.folder}, {currentfolder.name} );
numfolders = length(folder_list);
reference_table = cell(numfolders, 1);
Labels = cell(numfolders, 1);
for D = 1 : numfolders
    thisfolder = folder_list{D};
    [~, folderlabel, ~] = fileparts(thisfolder);
    finfo = dir(thisfolder);
    finfo([finfo.isdir]) = [];  %remove folders
    filenames = fullfile({finfo.folder}, {finfo.name});
    numfiles = length(filenames);
    feats_cell = cell(numfiles, 1);
    feats_labels = cell(numfiles, 1);
    for F = 1 : numfiles
        thisfile = filenames{F};
        thisdata = importdata(thisfile);
        feats_cell{F} = extractAccFeatures(thisdata);
        feats_labels{F} = folderlabel;
    end
    reference_table{K} = vertcat(feats_cell{:});
    Labels{D} = feats_labels;
end
reference_table = vertcat(reference_table{:});
Labels = vertcat(Labels{:});

Image Analyst 2022-9-3

在 MATLAB Online 中打开

Why 3-7? Are you trying to avoid . and .. which is what this does:

currentfolder(~[currentfolder.isdir]) = []; %remove non-folders

or are your first and second folder not . and ..?

Walter Roberson 2022-9-3

No, I cannot do that. Your current code is incorrect in design.

You should never rely on the relative position of files within a folder. None of the file systems you are likely to ever encounter impose any internal ordering on names, and none of the operating systems supported by MATLAB define a fixed order to return names in response to queries of directory information.

The NTFS file system commonly used by Windows has some "typical" behaviour, but the typical behaviour is not strictly adhered to, and is different from what most people expect... and depends in part on the region setting of the user who created the file system. So copying between drives can result in different ordering of files.

You should always code in terms of accepting or rejecting files based upon names or other attributes, not on the basis of order of files returned by dir()

请先登录，再进行评论。