This is cool but isn't it dangerous? Do we have a guarantee that the
memory registers n bytes after a data step character variable are safe to
change? Could this potentially overwrite the registers in another
applictaion or, worse, the operating system?
On Mon, 10 Dec 2001 15:13:21 -0500, Dorfman, Paul <Paul.Dorfman@BCBSFL.COM>
>That is exactly what I meant. Additionally, the same can be achieved as
>call poke (add, addr(tosubstr)+length(tosubstr), 4) ;
>Thus in a DATA step, a concatenation can always be coded in an alternative
>manner (it may be necessary to declare an extra variable in a LENGTH
>statement). Outside the DATA step, there are cases where a concatenation
>could hardly be replaced in a non-contrived way, in particular, when SQL is
>used to create concatenated lists, and where the concatenation is needed as
>a volatile expression only.
>Paul M. Dorfman