Help parsing a string

I need some help parsing a char array, hopefully someone has an idea. Matlab loads my table and the table header is not matlab compliant, so it creats a char array that is a property of the table column, and I need to parse that to get the proper table header name.

One of the char arrays is Original column heading: '(name='/mandySimulation_1/clock/simulationTime' type='DOUBLE' quantity='TIME' unit='s')'

So I need to create 4 variables, name, type, quantity, and unit. Each should have the subsequent value per the string. I tried using sscanf, but I can't figure out how to use it. Any suggestsion? I would like the name to keep the "/" characters because I will parse those later. It's a little easier because the / can be a delimiter.

The char arrays I'm working with are not the same length, but they always have the name, type, quantity, and unit variable name. Here is an example of another char array:

Original column heading: '(name='/mandySimulation_1/engine_1/longBlock_1/shortBlock_1/cylinderBlockAssembly_1/cylinder_1/ringLeakage_1/referenceDiameter' type='DOUBLE' quantity='LENGTH' unit='m')'

Thanks in advance for any help you can give.

Answer 1

[Ameer Hamza]

Try this

str = "Original column heading: '(name='/mandySimulation_1/clock/simulationTime' type='DOUBLE' quantity='TIME' unit='s')'";

name = regexp(str, '=''([^\s''.]*)''', 'tokens');

Result

>> name{1}
ans = 
    "/mandySimulation_1/clock/simulationTime"
>> name{2}
ans = 
    "DOUBLE"
>> name{3}
ans = 
    "TIME"
>> name{4}
ans = 
    "s"

It will work, as long as the strings are in the specified format.

[Russell Senior]

Works perfect. I thought it would involve the tokens flag, but regular and expressions and I don't really get along, I just don't use them enough. Thanks!

[Ameer Hamza]

I am glad to be of help.

[Russell Senior]

Of course, I find one that doesn't fit... Within my set of strings, I have at least one string that only has the name and type token. How do I extract the value by token name? The exact string that gave me an error is:

Original column heading: '(name='/mandySimulation_1/clock/stepCount' type='INTEGER')'

Notice it does not contain unit or quantity tokens.

So, I guess for this I would like to extract:

name = '/mandySimulation_1/clock/stepCount'

type = 'INTEGER'

quantity = ''

unit = ''

Thanks in advance.

[Ameer Hamza]

In this case, I think that struct is the most suitable way to organize the data. For example

str(1) = "Original column heading: '(name='/mandySimulation_1/clock/simulationTime' type='DOUBLE' quantity='TIME' unit='s')'";
str(2) = "Original column heading: '(name='/mandySimulation_1/clock/stepCount' type='INTEGER')'";

matches = regexp(str, '(\w*)=''([^\s''.]*)''', 'tokens');

S(numel(matches)) = struct('name', "", 'type', "", 'quantity', "", 'unit', "");
for i=1:numel(matches)
    match = matches{i};
    for j=1:numel(match)
        S(i).(match{j}(1)) = match{j}(2);
    end
end

Result:

>> S(1)
ans = 
  struct with fields:

        name: "/mandySimulation_1/clock/simulationTime"
        type: "DOUBLE"
    quantity: "TIME"
        unit: "s"
>> S(2)
ans = 
  struct with fields:

        name: "/mandySimulation_1/clock/stepCount"
        type: "INTEGER"
    quantity: ""
        unit: ""

Adapt it according to your requirement.

[Russell Senior]

Thanks so much! I hope wherever you are working, they are paying you well!

One minor correction:

S(i).(match{j}(1)) = match{j}(2);

Should be

S(i).(match{j}{1}) = match{j}{2};

Then it works 100% perfect. Nicely done, and thanks again!

[Ameer Hamza]

I am glad to be of help. Thanks for your kind remarks :)

In R2020a, both match{j}(1) and match{j}{1} works. Maybe this is something related to MATLAB releases.