Add columns to the tables that stored in a cell

Hey all, I have a 40 x 32 cell named Cnew. On the other hand, I have a 1 x 32 table named lat and 40 x 1 table named lon. I want to add corresponding latitude (lat) and longitude (lon) columns to each table in front of the existing column. So the problem is founding corresponding latitude and longitude based on the place of each table, and add latitude and longitude column for them. I know it's maybe complicated so I use this little example to tell what I need:

As I attached Cnew, at this time Cnew looks like this:

But in essence, they are placed based on latitude and longitude so in fact:

Now I have lon and lat for this data in lon and lat table that I attached. for example:

Now I have just one column in every T that represents a value. I want to add 2 columns in each T, one for latitude and one for longitude. For example, Orange T has lon=44.25 and lat=24.25. Blue T has lon=44.25 and lat=24.75. Green T has lon=44.25 and lat=22.25. Red T has lon=44.75 and lat=24.25. Now I want to add related lat and lon as columns Inside each T.

I don't know how to do this.

Thank You so much.

Answer 1

[Adam Danz]

"I want to add 2 columns in each T, one for latitude and one for longitude."

C2 is a 40x32 cell array of 336x1 tables. Is your goal to create a 40x32 cell array of 336x3 tables where the two additional columns are latitude and longitude values?

If that's what you're doing,

load('C2.mat')  %loads variable C2, a 40x32 cell of 336x1 tables
load('lat.mat') %loads variable lat, a 1x32 table
load('lon.mat') %loads variable lon, a 40x1 table

C2LatLon = cell(size(C2)); 
for i = 1:size(C2,2)            
    for j = 1:size(C2,1)
        n = height(C2{j,i}); 
        C2LatLon{j,i} = [repmat(lon(j,1),n,1), repmat(lat(1,i),n,1), C2{j,i}]; 
        C2LatLon{j,i}.Properties.VariableNames = {'lon','lat','data'};  % do you want to change the var names?
    end
end

The result, C2LatLon is a 40x32 cell array. Each element is a 336x3 table. An example is shown below.

head(C2LatLon{3})  %show first few rows of table #3
ans =
  8×3 table
     lon      lat     data 
    _____    _____    _____
    45.25    24.25     1.08
    45.25    24.25     8.88
    45.25    24.25    25.75
    45.25    24.25    35.46
    45.25    24.25     0.02
    45.25    24.25        0
    45.25    24.25     0.16
    45.25    24.25     0.11

Super table format

However, organizing the tables into a supertable would be more organized and would require less memory.

Here's a variant of Star Strider's deleted answer.

load('C2.mat')  %loads variable C2, a 40x32 cell of 336x1 tables
load('lat.mat') %loads variable lat, a 1x32 table
load('lon.mat') %loads variable lon, a 40x1 table

T = array2table(C2, 'VariableNames',  lat.Properties.VariableNames); 
T = [lat; T]; 
lon = [table(nan,'VariableNames',{'lon'}); lon]; 
T = [lon, T]; 

Here's a sample of what that looks like

T(1:8,1:6)
ans =
  8×6 table
     lon         lat1             lat2             lat3             lat4             lat5     
    _____    _____________    _____________    _____________    _____________    _____________
      NaN    {[    24.25]}    {[    24.75]}    {[    25.25]}    {[    25.75]}    {[    26.25]}
    44.25    {336×1 table}    {336×1 table}    {336×1 table}    {336×1 table}    {336×1 table}
    44.75    {336×1 table}    {336×1 table}    {336×1 table}    {336×1 table}    {336×1 table}
    45.25    {336×1 table}    {336×1 table}    {336×1 table}    {336×1 table}    {336×1 table}
    45.75    {336×1 table}    {336×1 table}    {336×1 table}    {336×1 table}    {336×1 table}
    46.25    {336×1 table}    {336×1 table}    {336×1 table}    {336×1 table}    {336×1 table}
    46.75    {336×1 table}    {336×1 table}    {336×1 table}    {336×1 table}    {336×1 table}
    47.25    {336×1 table}    {336×1 table}    {336×1 table}    {336×1 table}    {336×1 table}

Super Table 2 (Recommended)

This arrangement would be easiest to use. Replace row and column names with something more meaningful.

load('C2.mat') %loads variable C2, a 40x32 cell of 336x1 tables
load('lat.mat')%loads variable lat, a 1x32 table
load('lon.mat')%loads variable lon, a 40x1 table

[latMat, lonMat] = ndgrid(lat{:,:}, lon{:,:}); 
C = cellfun(@(T){T{:,:}}, C2); 
LLT = array2table([latMat(:).'; lonMat(:).'], 'RowNames', {'Lat','Lon'}); 
T = [LLT; array2table(cell2mat(C(:)'),'RowNames',compose('Data%d',1:height(C2{1})))];

Here's a sample of the first 8 rows and 10 columns.

T(1:8,1:10)
ans =
  8×10 table
             Var1     Var2     Var3     Var4     Var5     Var6     Var7     Var8     Var9     Var10
             _____    _____    _____    _____    _____    _____    _____    _____    _____    _____
    Lat      24.25    24.75    25.25    25.75    26.25    26.75    27.25    27.75    28.25    28.75
    Lon      44.25    44.25    44.25    44.25    44.25    44.25    44.25    44.25    44.25    44.25
    Data1     1.08     1.06     1.08     0.97     0.96     1.05        1     0.87     0.45        0
    Data2    39.79    20.96     8.88     3.83     3.03     2.55     0.42     0.67     1.46     2.73
    Data3     20.2    23.32    25.75    27.42    30.01    32.44    30.04    28.46    28.27    28.78
    Data4     31.9    34.19    35.46    35.01    32.79    30.68    32.92    31.81    23.53    10.19
    Data5        0        0     0.02     0.19     0.38     0.43     0.43     0.49     0.15        0
    Data6        0        0        0        0        0        0        0     0.01     0.04     0.08

[BN]

Thank you Both. Actually that's what I needed. Thank you both.

[Stephen23]

"organize the tables into a supertable would be more organized and would require less memory"

It would also be much simpler to work with. Splitting data up into lots of arrays usually makes it harder to work with (as you are finding out now). Just add a longitude and latitude variable to each table, then merge them all into one table (or even better, don't even split them up in the first place!), then you can process your data using the standard tools applied to that one table, e.g.:

https://www.mathworks.com/help/matlab/matlab_prog/grouping-variables-for-splitting-data.html

What you are doing now totally defeats the point of the table class.

[Star Strider]

@Stephen — That was the essence of my (deleted since it was not Accepted) Answer.