Plot multiple variables from table with different x axes

Hello!

I have a table where I have multiple (300+) different transect measurements, so I have columns named Distance_1, Ca_1, Fe_1, Distance_2, Ca_2, etc. I am wanting to plot multiple Ca columns against their corresponding Distance column but I cannot find a way to do it without doing each individually.

I can't figure out how I would loop through the columns since when I try to do something like

for i=1:10
    h=stairs(Map1.Distance_i,Map1.Ca_i);
    hold on
end

it says there is unrecognized table variable "Distance_i". I do have the columns in order, so it always goes Distance-Mn-Ca-Fe-Mg-XMg, so the next 'Distance' column is always n+6 but I run into the same issue with unrecognized table variables.

Thanks so much.

Answer 1

[Voss]

ncol = size(Map1,2);
for i=1:ncol/6
    h=stairs(Map1.(sprintf('Distance_%d',i)),Map1.(sprintf('Ca_%d',i)));
    hold on
end

Or:

ncol = size(Map1,2);
for i=1:ncol/6
    h=stairs(Map1{:,sprintf('Distance_%d',i)},Map1{:,sprintf('Ca_%d',i)});
    hold on
end

https://www.mathworks.com/help/matlab/matlab_prog/generate-field-names-from-variables.html

https://www.mathworks.com/help/matlab/ref/sprintf.html

https://www.mathworks.com/help/matlab/matlab_prog/access-data-in-a-table.html

[Hannah Cunningham]

Thanks so much!

[Voss]

You're welcome!

Answer 2

[Dyuman Joshi]

When you do this

h=stairs(Map1.Distance_i,Map1.Ca_i);

MATLAB looks for the column named Distance_i and Ca_i in the Map1, which do not exist, and thus gives the error you received.

To access the data, you can generate column names as strings and use them -

for i=1:10
t1=sprintf('Distance_%d\n',i);
t2=sprintf('Ca_%d',i);
h=stairs(Map{:,t1},Map{:,t2});
hold on
end

And since the columns in the table are arranged in an order, you can use indexing as well -

for i=1:10
    h=stairs(Map.(1+6*(i-1)),Map.(3+6*(i-1)));
end

Read here - Access Data in a Tables

[dpb]

Alternate table addressing syntax -- use variable name with .(x) where x is a variable holding the desired variable. The "." returns the whole referenced column without the need for the explicit 2D indexing and the curlies "{}".

@Dyuman Joshi used the same syntax with the the numeric indices, but the more involved one with the variable names...

hold on
for i=1:10
  D=compose('Distance_%d',i);
  C=compose('Ca_%d',i);
  stairs(Map.(D),Map.(C))
end

[Dyuman Joshi]

Force of habit ¯\_(ツ)_/¯

Answer 3

[David Hill]

figure;hold on;
for i=1:10
  stairs(eval(sprintf('Map1.Distance_%1d',i)),eval(sprintf('Map1.Ca_%1d',i)));
end

[dpb]

There's no need for eval here...just reference the table variable dynamically instead of the table itself---you have to code the variable name in somewhere, anyway...

Map1=array2table([[0:4]' rand(5,1)],'VariableNames',{'Distance_1','Ca_1'})

Map1 = 5×2 table
    Distance_1     Ca_1  
    __________    _______

        0         0.61311
        1         0.81437
        2         0.14077
        3         0.39314
        4         0.64307

i=1;
stairs(Map1.(sprintf('Distance_%1d',i)),Map1.(sprintf('Ca_%1d',i)));

Answer 4

[dpb]

There's also no need to loop; use the array inputs version of stairs

D_indices=[1:3:width(map1)];        % would be all; limit to what desired
C_indices=D_indices+1;              % Ca follows D in order
stairs(map1{:,D_indices},map1{:,C_indices})         % stairs with arrays X,Y

NOTA BENE: Above uses the indexing; could build the variable names(*), but the syntax with "dot" notation only works with a single variable so it doesn't help any.

(*) More efficient, simpler than rebuilding variable names from string operations is to simply extract them from the Properties.VariableNames of the table. For example, the above could be followed by

lengend(map1.Properties.VariableNames(C_indices))