Creating a 3D table (or equivalent)

Is it possible to create a 3D table to represent a stack of tables? I have N number of tables, each of which has the same number of columns and rows, the same column and row names, and the same type of data in each column. I'm looking for a way to stack them, so that each data set can be referenced by either it's string name, or the index in the stack.

You can create a table of tables like this:

initTable1 = table('Size',[6,4],'VariableTypes',{'string','double','double','double'})
initTable2 = table('Size',[6,4],'VariableTypes',{'string','double','double','double'})
initTable3 = table('Size',[6,4],'VariableTypes',{'string','double','double','double'})
stackTable = table(initTable1,initTable2,initTable3)

But I want to be able to define the number of tables that are included in 'stackTable' on the fly. How would I initialize such a table? The number of stacked tables will not be static (and will probably be on the order of 6-30 stacked tables). If it matters, the size of each individual table is relativly small.

I'm trying to do this because I'd like to be able to reference each table by either it's string name or by it's index, similar to how you can access each field in a traditional 2D table.

If this is not possible, or not advised, is there a better alternative?

Answer 1

[Ameer Hamza]

Instead of making tables of tables, I recommend making a cell array of tables because stacking makes more sense in that case. For example

initTable1 = table('Size',[6,4],'VariableTypes',{'string','double','double','double'});
initTable2 = table('Size',[6,4],'VariableTypes',{'string','double','double','double'});
initTable3 = table('Size',[6,4],'VariableTypes',{'string','double','double','double'});

stackTable = {initTable1, initTable2, initTable3};

Now you can get the data using the index of the table in stack and the variable name.

stackTable{index_in_stack}.variableName

[Peter]

Thanks, this works. I ended up adding another column to the cell array for name of table:

stackedData = cell(2,dataSetCount);
for i=1:dataSetCount
       stackedData{1,i} = dataSetName; %name concatenation not show, this is a unique name
       table = ImportDataTable() %input arguments not show, this gets a new data table
       stackedData{2,i} = table{:,:};
end
   
stackedData{1,:} %the names of the data sets (formated time, in my case)
stackedData{2,:} %the data

[Ameer Hamza]

Although it is a matter of convenience, if you also want to save names and tables, then I suggest using structs. Something like this

stackedData = repmat(struct('Name', '', 'Table', []), 1, dataSetCount);
for i=1:dataSetCount
       stackedData(i).Name = dataSetName; %name concatenation not show, this is a unique name
       table = ImportDataTable() %input arguments not show, this gets a new data table
       stackedData(i).Table = table{:,:};
end

[Peter]

It looks like when I assign the table to the fielf "Table" in the struct, that data is no longer represented as a table, but as a struct. Is that right?

So I can't call

stackedData(1).Table.NameOfCol1

Because stackedData(1).Table is no longer actually a table. That gives an error "Dot indexing is not supported for variables of this type". It has no longer has a field 'NameOfCol1'. Whereas if I keep a reference to table, I CAN still access each column (and extract it's data) by name in it:

table.NameOfCol1

[Ameer Hamza]

I guess this is happening because you are first extracting the content of the table and then assigning it to an element of the cell array. Change this line to

stackedData(i).Table = table{:,:};

change it to

stackedData(i).Table = table;

Also, I suggest changing the name of variables in these two lines from the table to something else. table is also the name of MATLAB's built-in function, it can cause an error later.

T = ImportDataTable() %input arguments not show, this gets a new data table
stackedData(i).Table = T;

[Peter]

That's it, thanks again!

[Ameer Hamza]

Glad to be of help.