FR: Don't error ppp when all coordinates are NA

[MichaelChirico]

I've got a situation where I'd like to feed all-NA coordinates to ppp (I have a loop where some subsets may be "empty" in the sense of not having any data).

The following works just as expected and accomplishes all I want (I later pixellate this and get the 0s out that I want)

ppp(xrange = 0:1, yrange = 0:1)

But this is an error:

ppp(x = NA_real_, y = NA_real_, xrange = 0:1, yrange = 0:1)

Error in ppp(x = NA_real_, y = NA_real_, xrange = 0:1, yrange = 0:1) :
all coordinate values are NA or NaN

I don't see a reason for the latter to error given the former works just fine (at worst, throw a warning).

Now, I need to throw a test into my code a la:

pixellate(if (all(is.na(xrange))) ppp(xrange = ...) else ppp(x = ....))

Which is considerably more dense/verbose/inscrutable

[rubak]

At the moment ppp doesn't allow any NA vaules. Can't you just filter out the NA values in your call to ppp?

x <- c(NA_real_, NA_real_)
y <- c(NA_real_, NA_real_)
ppp(na.omit(x), na.omit(y), xrange = 0:1, yrange = 0:1)

[baddstats]

I don't see a reason for the latter to error given the former works just fine

The two situations are completely different. NA is not the same as c() or numeric(0). In ppp(xrange, yrange) the coordinate vectors x and y have not been specified, so they default to numeric(0) yielding an empty point pattern. In ppp(NA, NA, xrange, yrange) the coordinates have been specified as a vector of length 1 containing NA. This is an error because ppp objects must not contain NA coordinate values (or a whole lot of code will break).

However, we could tweak ppp so that NA coordinate values are removed with a warning. I will discuss with the three amigos.

[MichaelChirico]

@baddstats thanks for the reply.

I understand the two are different -- this may be a case of a misleading error message, then.

See the edit -- the error I receive doesn't imply ppp can't handle missing values, but rather that the problem is there are no non-missing values. So in fact the error message is suggesting the source of the problem is exactly a numeric(0) issue.

Now I see that passing numeric(0) is not an error; thanks for correcting my understanding.

[baddstats]

Experimental code now inserted in ppp (in spatstat 1.49-0.004)

[baddstats]

For the record, ppp does not allow any coordinate values to be NA, NaN or infinite.
I will re-write the error message and the help file to make this abundantly clear.