How import file csv with vertical header

I am trying to import file csv. The header of the file is located on the vertical layout. For example, there are table with dimention 9x2, and the header is in column 1.

Name :; A
birth :; 22-Feb-1964
Age :; 1
Height :; 168.8 m
Weight :; 50.0 kg
Address :; C
School :; D
Note :; E
Range of time :; 07-May-14 13:46 - 12-May-14 08:12

Any of you have an idea how to import this csv file, by aware the type of each data (e.g.: double/string/etc)?

[Walter Roberson]

Please do not close a question that has an answer.

Answer 1

[Christopher Wallace]

Hi Agnes,

One way you could accomplish this is to first convert your data into an array and then transpose it.

a = readtable('YourCSV.csv', 'Delimiter', ';', 'ReadVariableNames', false);
b= table2array(a)'; % convert to array and transpose.

Then you'll need to convert it back to a table and update the variable names.

newTable = array2table(b(2,:));

To update the variables names you can use the command:

newTable.Properties.VariableNames = a{:,1};

BUT... the way the names are currently recorded in the data results in invalid table variable names due to the space and colon at the end.

You could remove those in a loop prior to setting them.

[Agnes Palit]

 opt = detectImportOptions(/User/file.csv);
 T = readtable(direction, opt);
    newT = table2array(T(:,:));
    newT = newT.';
    newT = array2table(newT(:,:));
    newT.Properties.VariableNames = {'Project', 'Subject', 'Measurement',...};
    output = newT(:,:);

I am using this code, and I think it adapts some of your code as well, but it does not solve the type of the data (e.g.: double/string, etc), do you have idea about this?

[Christopher Wallace]

What do you want the data types to be?

[Agnes Palit]

Name :; A
birth :; 22-Feb-1964
Age :; 1
Height :; 168.8 m
Weight :; 50.0 kg
Address :; C
School :; D
Note :; E
Range of time :; 07-May-14 13:46 - 12-May-14 08:12

according this example, there are string, datetime, double. Do you have idea how to adjust the type?

[Christopher Wallace]

You can use various methods to adjust the type but it depends on the starting type and what you want it to be. Without an example of what you want to do the best I can tell you is to look at:

str2double() and

num2str

[Agnes Palit]

Thank you, but I do not think the solution will be as simple as that. As specially to handle the type datetime. That is why I am using "detectImportOptions" to try handle the problem, and put code like this:

opt = setvartype(opt, [4 5], 'datetime');
 opt = setvaropts(opt, [4 5], 'InputFormat', 'dd-MMM yyyy', 'DatetimeFormat', 'eee uuuu-MM-dd HH:mm:ss' );

but this code just handle the file from "csv/another as long that is file (but with horisontal header not vertical)", not from "table"

Even we use function "DateTime" it will not solve the problem successfully. But if you have another idea which is more clear, please let me know. Thank you

[Christopher Wallace]

Please give me a specific example of what you're trying to do and all of the code you currently have.

[Agnes Palit]

Thank you, for the example data I already mention twice for you (above - that is the data I type manually which is actually put in csv file).

This is the following code:

filedir = '/Users/';
files1 = dir(fullfile(filedir, '*.csv'));
numFiles1 = length(files1);

for i = 1:numFiles1
  if (contains(files1(i).name, 'header') == 1)
      [header] = header(filedir, files1(i).name);
  end
end

this is the function "header.m":

function [output] = header(folder, name)
%UNTITLED Summary of this function goes here
%   Detailed explanation goes here

      direction = strcat (folder, name);
      opt = detectImportOptions(direction);

%     opt = setvartype(opt, [2 9], 'datetime');
%     opt = setvartype(opt, 4, 'double');
%     opt = setvartype(opt, 5, 'double');
%     opt = setvaropts(opt, [2 9], 'InputFormat', 'dd-MMM-yyyy', 'DatetimeFormat', 'eee uuuu-MM-dd HH:mm:ss' );
%     opt = setvaropts(opt, 4,'Suffixes',' m')
%     opt = setvaropts(opt, 5,'Suffixes',' kg')
%     T = readtable(direction, opt);
    T = readtable(direction, 'ReadVariableNames', false);
%     T = T.';
    newT = table2array(T(:,:));
    newT = newT.';
    newT = array2table(newT(:,:));
    newT.Properties.VariableNames = {'Name', 'birth', 'age', 'height', 'weight', 'address', 'school', 'note', 'time'};
    newT{2,1} = datetime(newT{2,1});
    output = newT(:,:);  
end

Answer 2

[Walter Roberson]

It is not possible to use readtable() for this purpose in any useful way, as readtable() expects the columns to be consistent data format, and would probably just give up and say that the second column was all character vectors.

You could perhaps adapt the code I posted in https://www.mathworks.com/matlabcentral/answers/285186-importing-data-without-knowing-number-of-columns#comment_368710 which goes through a bunch of trouble to figure out what the data types are of a cell array