I am trying to solve a system of nonlinear equations and compute the grouped mean of a dataset in MATLAB R2024b

Question:

My goal is to:

Solve the following nonlinear equations:

x2+y2−4=0

xy−1=0

Compute the grouped mean.

Generate two plots: a scatter plot for the solutions to the equations and a bar chart for the grouped mean values.

However, I am facing errors when trying to compute the grouped mean using varfun.

MATLAB throws an error in varfun, saying that "A grouping variable must be a column vector or a character array".

I need the grouped mean values for Value, but the function does not seem to work. How can I fix the error in varfun? Is it better to use grouptransform instead?

How can I correctly generate the two plots (scatter for equation solutions, bar for grouped means) without errors?

clc; clear; close all;
syms x y
eq1 = x^2 + y^2 - 4;
eq2 = x*y - 1;
sol = solve([eq1, eq2], [x, y]);
if isempty(sol.x)
error("No real solutions found!");
end
x_vals = double(sol.x);
y_vals = double(sol.y);
data = table([1; 2; 3; 4]', [10; 20; 30; 40]', 'VariableNames', ["ID", "Value"]);
groupedData = varfun(@mean, data, 'GroupingVariables', "ID", 'InputVariables', "Value");

Error using tabular/varfun (line 110)
A grouping variable must be a column vector or a character array.

fig = figure;
tiledlayout(2, 1);
nexttile
scatter(x_vals, y_vals, 'filled')
title("Solutions to the Nonlinear Equations")
xlabel("x values"), ylabel("y values")
nexttile
bar(groupedData.ID, groupedData.mean_Value, 'FaceColor', 'cyan')
title("Grouped Mean Values")
xlabel("ID"), ylabel("Mean Value")
exportgraphics(fig, "plot_output.png", "Resolution", 300);

Answer 1

[AR]

Hi @Likith,

The error is because MATLAB expects ID column to be a properly formatted column vector when used as a grouping variable in “varfun”.

Here, in the following implementation, “ID” is directly defined as a column vector:

data = table([1; 2; 3; 4], [10; 20; 30; 40], 'VariableNames', ["ID", "Value"]);

Grouped mean can then be computed and then plotted.

groupedData = varfun(@mean, data, 'GroupingVariables', "ID", 'InputVariables', "Value");
fig = figure;
tiledlayout(2, 1);
nexttile
scatter(x_vals, y_vals, 'filled')
title("Solutions to the Nonlinear Equations")
nexttile
bar(groupedData.ID, groupedData.mean_Value, 'FaceColor', 'cyan')
title("Grouped Mean Values")
exportgraphics(fig, "plot_output.png", "Resolution", 300);

The function “grouptransform” is useful when you want to retain the original table structure while adding a new column with computed values. It can be used as follows:

data.MeanValue = grouptransform(data, "ID", @mean, "Value");

You can refer the following MathWorks Documentation link for “grouptransform” function:

https://www.mathworks.com/help/matlab/ref/double.grouptransform.html

[Likith]

Thank you for clarifying about grouptransform as well. This helped!

Answer 2

[NVSL]

Hi @Likith

I understand you are trying to solve a couple of nonlinear equations.

I am able to reproduce the error occurred to you. As the error states, the input value of “Grouping Variables” should be a column vector.

Refer to the below documentation link for more details.

https://www.mathworks.com/help/matlab/ref/table.varfun.html#btyj4vl-1-GroupingVariables

On inspecting "data", “data[“ID”]” is a row vector. So, you can follow the below code to initialise data with entries as column vectors.

data = table([1; 2; 3; 4], [10; 20; 30; 40], 'VariableNames', ["ID", "Value"]); 

Hope it helps!

[Likith]

I wanted the plots and grouptransform as well but the other answer cleared everything up.Thanks