What does it mean when ~ is used as one of the outputs for a function?

I have seen a few scripts where ~ is used as one of the output variables for a function/command? What is this supposed to mean and how is it useful? For example, in one of the lines for the script for the function divergence.m, I find the line [px, ~] = gradient(u);

Answer 1

[James Tursa]

It means that the particular output that would ordinarily be returned to a variable at that spot is instead discarded automatically. E.g.,

[px, ~] = gradient(u);

is essentially equivalent to:

[px, dummy] = gradient(u); clear dummy

[Guillaume]

Well, in this particular case, it means that whoever wrote the line don't really know what they're doing. There's no point ignoring later outputs of the function*, simply do not ask for them. The line is exactly the same as:

 [px] = gradient(u).

The ~ is useful when you want to ignore earlier outputs of the function such as:

[~, py] = gradient(u);  %don't care about x gradient, but need second output.

*: unless the behaviour of the function changes depending on the number of output. This is not the case for gradient but is for find. e.g.:

idx = find(matrix); %return linear indices
[r, ~] = find(matrix);  %return rows, ignore columns

[James Tursa]

All good points. I will add that the ~ does not prevent the function from actually calculating that particular output argument. It is in fact calculated (the function still thinks it needs to return two outputs) ... it is just automatically discarded per the syntax. So using ~ is more of a coding convenience rather than a resource or time saver. (Is there a way for the function to detect when the caller has used ~ for an output?)

Answer 2

[LauraLee Austin]

It indicates that you are not using that output variable.

[FX,FY] = gradient(F)

Let's say you wanted the FY variable but not FX:

[~,FY] = gradient(F)