How do you store multiple variables?

Hello,

I would like to ask how you store multiple variables and control them.

Let's say there are 26 variables.

a=123;

b=..;

c=...;

......

z=123;

variable a,b,c,d are related and e,f,g,h are related and so on.

Would you make vector A=[a,b,c,d], B=[e,f,g,h] and so on and bring them out whenever you need like A(1) or A(2)?

I am wondering if I make vector and store variables in it, how will other people know what that numbers are when they read my code? what is A(1) or A(2)?

Should I put comment for every single variables what they are next to the vector?

I woud like to know the rule of thumb when you store the variables.

THANK YOU!!

[Matt J]

Why even assign separate symbols like a,b,c, when you have so many of them?

[David Hill]

Key is to index into either a cell array or matrix. Use cell array for different sizes.

m=[1 2 3 4
   5 6 7 8];
m(1,:)%would be first vector
m(2,:)%would be second vector
m={[1 2 3 4],[5 6],[7 8 9 10 11 12]};% or cell array for different sizes.

[dpb]

Create a separate variable only when it is the only thing it relates to -- if you have 3D acceleration values, don't make three variables, make an array.

If things are related but not the same, consider a struct of the variable with fields for the data associated.

Whatever you do, do not start out with A1, A2, A3, ... If do, the next thing you'll want to do is to address them programmatically and "there be dragons!"

[Stephen23]

"Would you make vector A=[a,b,c,d], B=[e,f,g,h] and so on..."

No, I would create one array for all of the data. Anything else will just make processing the data complex and inefficient.

"...and bring them out whenever you need like A(1) or A(2)?"

Yes.

"I am wondering if I make vector and store variables in it, how will other people know what that numbers are when they read my code? what is A(1) or A(2)?"

Using indexing is one of the basic, useful, and powerful aspects of MATLAB. If someone reads your code, then you can assume that they understand indexing.

[dpb]

And, as Walter points out below, you can identify key subscripts with IDs through other variables (too bad MATLAB doesn't have enumerations for such purposes) to label things.

The key there in my mind is to:

1. Don't go overboard in trying to name everything; just the key elements, and
2. Use meaningful but still short names so code isn't cluttered by long names making even harder to read than just the integers.

[Duckyoon Go]

Thank you so much to everybody!! I got the idea!

Answer 1

[Walter Roberson]

Your code is not required to use numeric indices in the body of the code.

pitch = 1; yaw = 2;

...
cos(A(pitch))*sin(A(yaw))

[dpb]

I once had to maintain a production, real-time instrumentation code built around an indexing system of arrays. The original idea was one could make coding changes by changing the indexing variables. It never worked out that way in the end, but couldn't convince employer to rewrite.

It ended up with code that looked like--

FUNCTION ConvergeSelectedCentroid%
    ConvergeROI% = Calib!(228)
    DIM CalibTime AS TimeStruct, NormalTime AS TimeStruct
    CalibTime.secs = SysCon!(132)
    NormalTime.secs = SysCon!(131)

    ' Find the Carbon Escape centroid and then try to align to desired
    AdjustedPeak% = FALSE
    MaxIterations% = Calib!(219)
    FOR i% = 1 TO MaxIterations%
        Origin% = ConvergeROI% * 20
        DeltaChannel! = Calib!(Origin%+20) * 100! - Calib!(Origin%+38)
        IF ABS(DeltaChannel!)<Calib!(214) THEN
            AdjustedPeak% = TRUE
            EXIT FOR
        END IF
        Calib!(211) = Calib!(211) + DeltaChannel! * SysCon!(243) * SysCon!(244)
        HVSet
        ...

And, more even more cryptic.

Calib!() and SysCon!() were global arrays.

It, of course, wasn't written in MATLAB... :(