Raw Message

Message-ID: <200305022014.h42KEmQm004425@erdos.math.unb.ca>
Date: 2003-05-02T20:14:48Z
From: Rolf Turner
Subject: Creating Dummy Variables with if else phrase

There is a bit of a trap for young players involved, but:

	> x <- NA
	> is.real(x)
	[1] FALSE
	> x <- c(1,NA)
	> is.real(x[2])
	[1] TRUE
	> x <- as.real(NA)
	> is.real(x)
	[1] TRUE
	> typeof(x)
	[1] "double"
	> x <- NA
	> typeof(x)
	[1] "logical"

It makes sense when you think it through:  An ``NA'' could be
anything on its own.  When it's an entry of a vector with real
entries then it oughta be a ``not available real''.  It's a bit
subtle when you first bump into it, but it all clicks into place.

Be that as it may --- to avoid the trap, use is.na() if you want
to check whether a value is missing.

				cheers,

					Rolf Turner