Can you please explain why do we do this code?

%Get antenna pairs and convert to matlab matrix

ants_pairs = vtrigU.vtrigU.GetAntennaPairs(mySettings.txMode);

TxRxPairs = zeros(ants_pairs.Length,2);

for ii = 1: ants_pairs.Length

TxRxPairs(ii,1) = double(ants_pairs(ii).tx);

TxRxPairs(ii,2) = double(ants_pairs(ii).rx);

end

Answer 1

[dpb]

To illustrate how NOT to write MATLAB code, apparently...

The person responsible for this wanted a Nx2 vector of the two members of the stuct variable; what use is made of them later is up to whatever the unshown code does later on.

But, the same end is obtained "the MATLAB way" without a for..end loop and doesn't need the preallocation step, either --

%Get antenna pairs and convert to matlab matrix
ants_pairs = vtrigU.vtrigU.GetAntennaPairs(mySettings.txMode);
TxRxPairs=double([ants_pairs.tx ants_pairs.rx]);

The double is needed only if the members in the struct are some other data type and then may still not be any need to change them from their default data type; again there's no way to judge the rest of the code.

[Walter Roberson]

TxRxPairs=double([ants_pairs.tx(:), ants_pairs.rx(:)]);

since we do not know their orientation.

[Ashis Pandey]

Why do we say double instead of single or why double specifically?

[dpb]

Can't tell without knowing what the rest of the code does with those values -- as noted, it may not make any difference at all; I don't know that the values in the original data struct aren't already double.

The default data type in MATLAB is the double; historically the other numeric classes were supported minimally if at all with math operations. It's possible this code dates from some time "way back when" before the other numeric types were more widely supported as now.

But, that's all conjecture and there's nothing so far that gives any hints as to there being a "why". With the initial code, the author clearly was not experienced in MATLAB so it may be only another symptom of that inexperience.

[Ashis Pandey]

Thankyou very much you seem very helpful, I really want to understand the code, I can send you the full info if you have the time.

Please let me know if possible, I can arrange any possible help like fund from my side to accommodate your time

Answer 2

[Walter Roberson]

The manufacturer's example code uses double() at that point, but does not explain why.

When using ActiveX controls (or when using Java methods), it is common that return values are in some internal array format that makes sense to the ActiveX control, and that the control supplies a double() method to convert to MATLAB double array.

For example this particular ActiveX control is written in C#. When it returns a C# array to MATLAB, it might be a header together with a block of memory in Column Major order, or in Row Major order, or it might be a header and a List, each element of which is a List of numbers that is a row or a column... all kinds of different ways that a program might store an array of data. The double() is not necessarily being used to convert between (for example) unsigned 16 bit integer and double precision: the data might already be double precision internally but arranged in a different way than MATLAB arranges data, so double() might be there to signal conversion of internal array format.