[Rd] Minor issue in code of 'diffinv.vector' in R 2.15.2

Martin Maechler maechler at stat.math.ethz.ch
Mon Jan 28 11:44:58 CET 2013


>>>>> Suharto Anggono Suharto Anggono <suharto_anggono at yahoo.com>
>>>>>     on Sun, 27 Jan 2013 23:54:50 -0800 writes:

    > In R 2.15.2 (and not before), in the definition of
    > function 'diffinv.vector' in package stats, there is
    > difference <- as.integer(differences)

    > I believe

    > differences <- as.integer(differences)

    > is intended, because 'difference' is not referenced
    > anywhere. However, without conversion of 'differences' to
    > integer, 'diffinv.vector' in R 2.15.2 works OK.

Thank you, Suharto,
you are right... about the typo above, the tweaks below,
*and*  ... ;-) ... that it is a minor issue

  -> changed for R-patched and R-devel.

Martin Maechler, ETH Zurich

    > Also, to make 'differences' consistently integer,

    >         diffinv.vector(diffinv.vector(x, lag,
    > differences-1, diff(xi, lag=lag, differences=1)), lag, 1,
    > xi[1L:lag])

    > in the end part can be changed to

    >         diffinv.vector(diffinv.vector(x, lag,
    > differences-1L, diff(xi, lag=lag, differences=1L)), lag,
    > 1L, xi[1L:lag])



More information about the R-devel mailing list