How to extract coefficients of custom variables in symbolic expression in MATLAB?

I want to extract the coefficient of i_B from the attached symbolic experession called omega_B and call it p for example, how to do that?

[Dyuman Joshi]

Hello @FAISAL ALQARNI, can you post the whole code?

[FAISAL ALQARNI]

I found a way to achieve this using Symbolic substitution subs

Since the expression is already factored in i_B, j_B and k_B, the coefficient of any of these symbolic variables can be simply extracted by setting its value to 1 and setting the other variables to 0 using subs

Example:

Extracting the coefficient of i_B from a symbolic expression called omega_B and call it p

syms psi phi theta i_B j_B k_B psi_dot phi_dot theta_dot
omega_B = (phi_dot - psi_dot*sin(theta))*i_B + (theta_dot*cos(phi) + psi_dot*cos(theta)*sin(phi))*j_B + (psi_dot*cos(phi)*cos(theta) - theta_dot*sin(phi))*k_B

omega_B = 

p = subs(omega_B,[i_B,j_B,k_B],[1,0,0])

p = 

Please let me know if there is a better approach.

Answer 1

[Paul]

syms psi phi theta i_B j_B k_B psi_dot phi_dot theta_dot
omega_B = (phi_dot - psi_dot*sin(theta))*i_B + (theta_dot*cos(phi) + psi_dot*cos(theta)*sin(phi))*j_B + (psi_dot*cos(phi)*cos(theta) - theta_dot*sin(phi))*k_B

omega_B = 

[c,t] = coeffs(omega_B,i_B)

c = 

t = 

p = c(t == i_B)

p = 

Get all the components simultaneously. Using both output arguments of coeffs forces the first output argument to be in the desired order (at least I think that's how it works)

clear c t
[omega,t] = coeffs(omega_B,[i_B j_B k_B])

omega = 

t = 

p = omega(1); q = omega(2); r = omega(3); % if individual components are needed as individual variables
[p;q;r]

ans = 

[p,q,r]=struct('omega',num2cell(omega)).omega  % one line approach

p = 

q = 

r = 

Answer 2

[Chunru]

Let iB=1, jB=0 and kB=0. Then compute omegaB.

[FAISAL ALQARNI]

I couldn't find a function called let, if you meant setting the variable by simply typing:

i_B=1; j_B=0; k_B=0;

Then this, unfortunately, won't work 😬.

syms psi phi theta i_B j_B k_B psi_dot phi_dot theta_dot
omega_B = (phi_dot - psi_dot*sin(theta))*i_B + (theta_dot*cos(phi) + psi_dot*cos(theta)*sin(phi))*j_B + (psi_dot*cos(phi)*cos(theta) - theta_dot*sin(phi))*k_B

omega_B = 

i_B=1,j_B=0,k_B=0

i_B = 1

j_B = 0

k_B = 0

p = omega_B

p = 

[Dyuman Joshi]

Not a function let, but the english word 'let'.

He suggested the exact thing you found out on your own (what you mentioned in a comment above).

Using subs to substitute and obtain value is what Chunru meant by computing.

[Chunru]

Exactly.