converting a table of numeric data into a double array?
515 views (last 30 days)
Show older comments
Hi,
I am trying to convert a table of numeric data to a double array using some of the methods presented before in some answers to similar questions like:
1- table2array function
2- X{:,:}
But they both don't work for me. They change the table into a cell, not a double array
0 Comments
Accepted Answer
BN
on 8 Apr 2020
I believe it's mean that you have non-numeric data in your table.
Try this:
T_cell = table2array(T);
T_array = cell2mat(T_cell); % T_array is output
% if it is single so after this two lines use:
% T_array = double(T_array)
3 Comments
BN
on 8 Apr 2020
I believe it happens because your first data table includes text and numeric data.
Do you have a problem with using this?
data_numNEW = table2array(data_num );
data_numNEW = cell2mat(data_numNEW ); % T_array is output
I do not see your data but if the problem is you have char instead of number this maybe helps you:
data_numNEW = table2array(data_num );
data_numNEW = cellfun(@double,data_numNEW,'uni',0)
More Answers (1)
Chhanda
on 19 Oct 2023
Edited: Walter Roberson
on 19 Oct 2023
My code is this :
data=readtable('MainData.xlsx');
X1= data(:, 3);% Assuming time, temperature, humidity are columns 3, 4, and 5
X2=data(:, 4);
X3=data(:, 5);
% Convert variables to tables
X1Table = table(X1);
X2Table = table(X2);
X3Table = table(X3);
% Extract the dependent variable (soil moisture)
Y = data(:, 6); % Assuming soil moisture is in column 6
n=height(X1);
onesTable = table(ones(n,1));
a = [onesTable, X1Table, X2Table, X3Table];
c=pinv(a)*Y;
I am getting this error:
Error using svd
First input must be single or double.
Error in pinv (line 18)
[U,s,V] = svd(A,'econ','vector');
Error in Project1 (line 21)
c=pinv(a)*Y;
Kindly suggest a solution.I think i need to change my table to double format ..but unable to do so using table2array().
Kindly help.
1 Comment
Stephen23
on 19 Oct 2023
Edited: Stephen23
on 19 Oct 2023
"Kindly suggest a solution."
Use numeric arrays for storing basic numeric data (not lots of superfluous nested tables).
Use curly-brace indexing to access table content:
The PINV documentation clearly states that its input must be SINGLE or DOUBLE (i.e. numeric). Instead of doing that, you are providing it with nested tables.
"Convert variables to tables": they are already tables, why nest them inside more tables? Why do you need so many tables? Lets try it without all of those tables, e.g. by using curly-brace indexing:
T = array2table(rand(7,6)); % fake data alternative to READTABLE
Y = T{:,6};
n = height(T);
a = [ones(n,1), T{:,3:5}];
c = pinv(a)*Y % your approach
c = lsqminnorm(a,Y) % the recommended approach
See Also
Categories
Find more on Tables in Help Center and File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!