MATLAB: Is it possible to stack tables

MATLABtables

Hi everyone, I'm facing the following problem: I do several testruns. There are about 10 tests per run, which results in 10 .csv files. I read them with Matlab, do some calculations and store about 6 values per test which I write in a table, thereby creating a 6×10 table. Those values are sometimes strings and sometimes numbers. Then I change the setting and do the next test run, evaluate the files and create the next table. I end up with about 8 tables. Now I would like to somehow stack those tables in order to compare (and plot) the influence of the setting on my test results. I would like to do something like:

all_tables(1, :, 🙂 = table1;

all_tables(2, :, : ) = table2;

etc.

and then read the values back, e.g.

val1 = all_tables(:, 2, 1) ;

Is that possible or is there any other format which is easier to handle than tables?

Thank you in advance 🙂

Best Answer

As far as I can tell, tables are limited to two dimensions. Here are some ideas for storing an arbitrary set of data of arbitrary sizes:

It is possible to concatenate tables (of course this requires that the columns/variables be of the same class). Simply add a new index variable, which tells you which test the data is from.
You could put the tables into a cell array. This might be the easiest solution.
You could avoid the tables entirely, and just use a 3D cell array. This will be difficult if each test produces different length outputs.

Related Solutions

MATLAB: Formatting more than 600 table

A simple code (which does not preserve completely the table header):

tabledir = 'C:\wherever\the\files\are\on your\drive';
tablecount = 600; %how however many there are
tableout = 'alltables.txt' %or however you want to name it
tables = cell(tablecount, 1);
%read tables one by one into cell array
for tableidx = 1:tablecount
   tablefile = sprintf('table%02d.txt', tableidx);
   if ~exist(fullfile(tabledir, tablefile), 'file')
      warning('skipping table %s. file does not exist', tablefile);
   else
      tables{tableidx} = readtable(fullfile(tabledir, tablefile)); %read table
   end
end
alltables = vertcat(tables{:}); %concatenate all tables
writetable(alltables, fullfile(tabledir, tableout));

Note, that I've not tested nor debugged it, there may be some small typos, syntax errors.

MATLAB: Concatenating Tables with Regionprops

Letting your table grow dynamically might not an efficient coding approach. It would be better if you save the tables in a cell array and then attach them together after the loops are completed. For example

regions = cell(1,n)
for i = 1:n
    regions{i} = regionprops('table', ..)
end
all_regions = vertcat(regions{:});

Also, this approach does not require you to manually initialze the table.

Best Answer

Related Solutions

MATLAB: Formatting more than 600 table

MATLAB: Concatenating Tables with Regionprops

Related Question