How do I horizontally concatenate a cell of strings with a table of numbers?

Code:

function [finalTable] = featureExtractionTrainFileWise(varargin)

for masterVariable=1:nargin;%Goes over all the files

    %Step 1: Load the data for one file into tables

    windowWidth=64;%vary this to vary the size of the window
    windowOverlap=0.5;
    windowStep=windowWidth*windowOverlap;
    table_x1 = [];
    table_y1 = [];
    table_z1 = [];
    table_class1 = [];
    table_accel1 = [];

    t = readtable(varargin{masterVariable});
    table_x1 =  t.X_mG_;%contains numbers
    table_y1 =  t.Y_mG_;%contains numbers
    table_z1 =  t.Z_mG_;%contains numbers
    table_class1 =  t.Classification; %contains strings

    [rows,~] = size(table_x1);

    %Step 1a: Calculate acceleration for the data
    table_accel1=zeros(rows,1); %Pre-allocate an array of all zeroes to save compilation time.
    for i=1:rows
        vector = [table_x1(i,1) table_y1(i,1) table_z1(i,1)];
        table_accel1(i,1) = norm(vector);
    end

    %remove any extra rows (that won't fit in one window)
    extraRows = mod(rows,windowWidth);
    firstExtraRow = rows-extraRows +1;
    table_x1(firstExtraRow:rows,:)=[];
    table_y1(firstExtraRow:rows,:)=[];
    table_z1(firstExtraRow:rows,:)=[];
    table_accel1(firstExtraRow:rows,:)=[];
    %Step 2: Slide a window down the columns and calculate the mean and
    %standard deviation of each window (for each column)

    numWindows = ceil(rows/windowStep);%num windows is the number of windows we will have, based on 
    %the number of rows we have, the amount of overlap, etc.

    table_xMean = zeros(numWindows,1);
    table_xStDev = zeros(numWindows,1);
    table_yMean = zeros(numWindows,1);
    table_yStDev = zeros(numWindows,1);
    table_zMean = zeros(numWindows,1);
    table_zStDev = zeros(numWindows,1);
    table_accelMean = zeros(numWindows,1);
    table_accelStDev = zeros(numWindows,1);

    i=1;
    while i<rows;
        currentRow = floor(i/windowStep) + 1; %Returns 1 the first pass through the loop, 2 the second pass, 3 the third,etc.
        firstPoint=i*windowStep;
        lastPoint=i+windowWidth;
        table_xMean(currentRow,1) = mean(table_x1(firstPoint:lastPoint,1),1); %average accel_x for the window
        table_xStDev(currentRow,1) = std(table_x1(firstPoint:lastPoint,1),[],1); %standard deviation of accel_x for the window
        table_yMean(currentRow,1) = mean(table_y1(firstPoint:lastPoint,1),1);%average accel_y for the window
        table_yStDev(currentRow,1) = std(table_y1(firstPoint:lastPoint,1),[],1);%standard deviation of accel_y for the window
        table_zMean(currentRow,1) = mean(table_z1(firstPoint:lastPoint,1),1);%average accel_z for the window
        table_zStDev(currentRow,1) = std(table_z1(firstPoint:lastPoint,1),[],1);%standard deviation of accel_z for the window
        table_accelMean(currentRow,1) = mean(table_accel1(firstPoint:lastPoint,1),1);%average accelartion for the window
        table_accelStDev(currentRow,1) = std(table_accel1(firstPoint:lastPoint,1),[],1);%standard dev. of accel for the window
        i = i + windowStep; 
    end

    [number,~]=size(table_xMean);
    table_class1(1:number,:)=[];

    finalTable1=[table_xMean,table_xStDev,table_yMean,table_yStDev,table_zMean,table_zStDev,table_accelMean,table_accelStDev,table_class1];
    finalTable=[finalTable;finalTable1];

 end
 end

Error:

 Error using horzcat
 Dimensions of matrices being concatenated are not consistent.

 Error in featureExtractionTrainFileWise (line 98)
    finalTable1=[table_xMean,table_xStDev,table_yMean,table_yStDev,table_zMean,table_zStDev,table_accelMean,table_accelStDev,table_class1];

If I remove the "table_class1" from the horizontal concatenation statement, so it would read:

finalTable1=[table_xMean,table_xStDev,table_yMean,table_yStDev,table_zMean,table_zStDev,table_accelMean,table_accelStDev];

then the error goes away, indicating that it is table_class1 that has inappropriate dimensions.

However, if my understanding is correct, it should have the same size as table_xMean, from:

    [number,~]=size(table_xMean);
    table_class1(1:number,:)=[];

I checked for off by one errors by trying:

    table_class1(1:number+1,:)=[];

and

    table_class1(1:number-1,:)=[];

But neither of those fixed it.

I suspect the problem has something to do with the fact that table_class1 is a cell array instead of a matrix (as dispcell(table_class1) prints it's output). I started using MATLAB today, so I'm not sure how to fix this, or if it's even a problem.

What can I do to concatenate table_class1 against the rest of the tables?

[Azzi Abdelmalek]

Can you post a short example with expected result?

[James Stevens]

Short example input:

456  256  -848  Jumping Jacks
728  224  -2096  Jumping Jacks
1392  400  -1928  Jumping Jacks
2296  1008  -2328  Jumping Jacks
3168  1152  -2224  Jumping Jacks
3120  1640  -1184  Jumping Jacks
2608  1584  -408  Jumping Jacks

Expected result:

[#xAvg #xStDev #yAvg #yStDev #zAvg #zStDev #accelAvg #accelStDev "Jumping Jacks"]

(the #xAvg would contain whatever the numerical average of the x column (for that window) is).

If there was more input it would take several windows to cross all the input, so it would output several rows of that form.

Answer 1

[Madhav Rajan]

I undestand that you want to concatenate a cell of strings with a table of numbers. Looking at the script that you have provided, it appears that the following line:

   finalTable1=[table_xMean,table_xStDev,table_yMean,table_yStDev,table_zMean,table_zStDev,table_accelMean,table_accelStDev,table_class1];

concatenates a cell array string with other vectors to form a matrix. It is not possible to concatenate a vector and a cell array of strings and form a matrix because all the elements must be of the same type. Suppose you have 'Age' and 'LastName' as two variables, one is a vector of numbers and the other is a cell array of strings.

Age = [38;43;38;40;49];
LastName = {'Smith';'Johnson';'Williams';'Jones';'Brown'};

If you concatenate them in the sameway as you did to obtain finalTable1, you would get the same error:

finalTable1 = [Age LastName]

Error using horzcat Dimensions of matrices being concatenated are not consistent.

Instead you can create a table out of these elements and assign that to finalTable1 as shown:

finalTable1 = table(Age, LastName);

Hope this helps.

[polar]

I have this problem: I wrote a script for downloading a 4 stocks price's vector that i called "tickers". In my script I display a 4x4 matrix of all correlations. What do I have to do for displaying all tickers names together with the data? Thanks