Once more... Avoid global variable!

114 views (last 30 days)
Rafael Kübler
Rafael Kübler on 29 Nov 2017
Commented: Rafael Kübler on 30 Nov 2017
Hello together,
I'm writing many function which i often use during a matlab session. To make them a little more compfortalbe to use i'd like to use a data structure in some way, where i can store some preference data. The preference data is for example some input arguments the functions have in common. But because the values of these input arguments change from time to time, so i can not set it as default values in the function.
Global variables look perfect for this task, but i've read a lot of comments that global variables are bad and i'd rather avoid using them. But what other ways are there, to store my preference data and use it in different functions.
thank you for your help.
Rafael
  1 Comment
Jan
Jan on 29 Nov 2017
This is a good question, +1. You state "Global variables look perfect for this task". The magic core in this formulation is "look": In fact, globals look like they solve the problems and they really do so - but in a long term view they will cause even more troubles, which can grow to a such severe level, that the code cannot be debugged or maintained anymore.

Sign in to comment.

Accepted Answer

Jan
Jan on 29 Nov 2017
You can create a "global function" to store static parameters:
function G = theGlobalData()
G.pi = 3.14;
G.imag = 1i;
G.PIN = 5553;
end
You might add some input arguments also or store G persistently:
function G = theGlobalData(Name, Value)
persistent G_
if isempty(G_)
G_.pi = 3.14;
G_.imag = 1i;
G_.PIN = 5553;
end
if nargin > 0
G_.(Name) = Value;
end
G = G_;
end
This shares the problem of global variables, that e.g. running multiple instances of your program can cause collisions of the values. But at least you can debug each access of this function. Setting a breakpoint in it allows to track all events.
I'd prefer sharing the struct inside the code by using input and output arguments. Passing an additional argument is neither clumsy nor slow and there is no danger of interferences with other instances of your code or other software.
These ideas are "create a parameters function" and "pass a struct" of Stephen's answer.
  7 Comments
Rafael Kübler
Rafael Kübler on 30 Nov 2017
Oh, yes it works however now.

Sign in to comment.

More Answers (3)

Image Analyst
Image Analyst on 29 Nov 2017

Jos (10584)
Jos (10584) on 29 Nov 2017
Take a look at the setappdata and getappdata. You can use the root handle to store variables.
setappdata(0,'MyVar',1:10)

per isakson
per isakson on 29 Nov 2017
Edited: per isakson on 29 Nov 2017
Download parse_pv_pairs by John D'Errico and run this example
preference.Viscosity = 1;
preference.Volume = 1;
preference.Pie = 3.141592653589793;
preference.Description = 'test';
params = examplefun( preference, 'Viscosity',2, 'Volume',3 );
and inspect params
>> params
params =
Viscosity: 2
Volume: 3
Pie: 3.1416
Description: 'test'
where
function ip = examplefun( nv, varargin )
ip = parse_pv_pairs( nv, varargin );
end
In examplefun you refer to the input arguments as ip.Viscosity, ip.Volume, ip.Pie and ip.Description.
This example doesn't fully meet your requirements, since the structure, preference, must contain all fields.
This solution may be developed further. Download CATSTRUCT, by Jos (10584) and run this modified example
preference.Viscosity = 1;
preference.Pie = 3.141592653589793;
params = examplefun( preference, 'Viscosity',2 );
and inspect params carefully
params =
Description: 'Default values'
Pie: 3.1416
Viscosity: 2
Volume: 12
>>
where
function ip = examplefun( nv, varargin )
warning('off','catstruct:DuplicatesFound')
default.Viscosity = 11;
default.Volume = 12;
default.Pie = 3.141592653589793;
default.Description = 'Default values';
nv = catstruct( default, nv );
ip = parse_pv_pairs( nv, varargin );
end
Both these FEX-contributions are good.
  1 Comment
Rafael Kübler
Rafael Kübler on 29 Nov 2017
Than you very much Per for your answer, but i think i stick with Jans' solution.

Sign in to comment.

Categories

Find more on Debugging and Analysis in Help Center and File Exchange

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!