Sort columns of a table by the value corresponding

Ive attached a sc of my table, although it continues to 86 columns, I would like to sort the values corresponding to each variable in decending order, and have the variable names be sorted along with the value.

The table dimensions is 1x86 as seen in the screenshot

[the cyclist]

I don't really follow what you want. Do you want each column sorted independently from the others?

And then you want the columns sorted alphabetically according to the column names (e.g. Ac first, Zr last)?

Out of curiosity ... why? Guessing homework.

[Turner]

Row 1 sorted in descending order. I'm able to do that if I just take row 1 and make it an array and use sort(), but then the names of each column are no longer associated with the correct numerical value, so I can't label a figure correctly. So basically I'm trying to get the row sorted in decending order while its still a table so each value is aligned with its correct variable name.

An open ended project, but yes hw essentially.

[Stephen23]

If the table was arranged transposed, then this would be trivial using SORTROWS():

V = [0.141;0;0.0146;0.0045;0.0567];
T  = table(V,'RowNames',{'H','He','Li','Be','B'})

T = 5×1 table
            V   
          ______

    H      0.141
    He         0
    Li    0.0146
    Be    0.0045
    B     0.0567

T = sortrows(T,'V')

T = 5×1 table
            V   
          ______

    He         0
    Be    0.0045
    Li    0.0146
    B     0.0567
    H      0.141

Better data design makes code simpler, more robust, and more efficient.

Answer 1

[Walter Roberson]

 [~, idx] = sort(table2array(YourTable), 2, 'descend');
  out = YourTable(:, idx) ;

Note that this implementation will only work for a single row, as it rearranges the table object. If you needed multiple rows then the variable names would have to be mixed with the numeric values, which is possible with a change of representation.

[Turner]

Which representation would be the best to change it to in order to keep the variable names associated with each value? As that is what I'm attempting

[Walter Roberson]

My proposal is the same as what Voss suggested, just that I tend to use different helper functions, such as

[~,idx] = sort(table2array(T),'descend');

The crucial step is taking the sort order information from sort() and using it to index the table variables, which re-arranges the table to have the variables in that order.

The "different representation" I was referring to would apply to the case where you had multiple rows in the table and wanted to process each row individually: in such a case you would not be able to have two different variable orders in two different rows of the same table() object. In such a situation you could use representations such as

'Li', 0.146, 'H', 0.141, 'He', 0
'H',  0.314, 'Li', 0.11', 'C', 0.04

right inside the table, with odd-numbered variables being element abbreviations and the even-numbered variables being the coefficient that went with it.

Or you could use the data to construct a table such as

'Li: 0.146', 'H: 0.141', 'He: 0'
'H: 0.314',  'Li: 0.11', 'C: 0.04'

with text for all entries. Or you might prefer text but

'0.146: Li', '0.141: H', '0: He'
'0.314: H',  '0.11: Li', '0.04: C'

The method that I had suggested, and which Voss showed explicit code for, works fine for a single row; you only need to care about alternate forms if you have multiple rows.

Answer 2

[Voss]

% I make a table of random data with 10 columns and 1 row, to represent your table
C = num2cell(rand(1,10));
T = table(C{:},'VariableNames',cellstr(('A':'J').'))

T = 1×10 table
       A         B          C          D          E          F         G          H         I          J   
    _______    ______    _______    _______    _______    _______    ______    _______    ______    _______

    0.87544    0.0112    0.93254    0.65738    0.71669    0.62776    0.3965    0.90837    0.8422    0.10215

% sort the values, descending
[~,idx] = sort(T{:,:},'descend');

% reorder the table columns
T = T(:,idx)

T = 1×10 table
       C          H          A         I          E          D          F         G          J         B   
    _______    _______    _______    ______    _______    _______    _______    ______    _______    ______

    0.93254    0.90837    0.87544    0.8422    0.71669    0.65738    0.62776    0.3965    0.10215    0.0112

[Turner]

This works, thank you!

it's currently not allowing me to accept the answer but I will continue trying

[Ron]

I have this data. the last row (the one highlighted in blue) is the sum of all the values in each column. I want to re-arrange all the columns as per the descending values in this last row but I am unable to do so. I tried your code and it did work for all other tables but I dont know why its not working for my data. Can you please help?

[Walter Roberson]

@Ron

sortrows(Data, 9, 'descending')

[Ron]

Thankyou so much for taking out time and replying to my question. Although I was able to find out that one of the elements was a string instead of numeric and thats why I was not able to sort to but thankyou once again for helping.