[R] Odp: How to change row value based on previous row value

[Petr PIKAL]

Peter Dalgaard <P.Dalgaard at biostat.ku.dk> napsal dne 08.10.2007 16:31:28:

> Petr PIKAL wrote:
> > Hi
> >
> > r-help-bounces at r-project.org napsal dne 08.10.2007 14:24:13:
> >
> > 
> >> Hi,
> >>
> >> I would like to fix the data in the following data.frame, and I am
> >> having trouble coding this in R - help please!
> >>
> >> 
> >>> x
> >>> 
> >>   A    B   x    y     z
> >> 1 1 10.0 100 1000 10000
> >> 2 2 19.8 200 2000 20000
> >> 3 3 20.1 300 3000 30000
> >> 4 4 20.3 400 4000 40000
> >> 5 5 30.0 500 5000 50000
> >>
> >> Column B is the problem.
> >>
> >> The logic I need to apply (this is in pseudo code ... I don't know 
how
> >> to do this in R)
> >>
> >> for all rows, r
> >> if x$B[r] <= x$B[r-1] + a 
> >>    then x$B[r] = x$B[r-1]
> >>
> >> i.e. If a=1, then I would end up with the data.frame
> >>
> >> 
> >>> x
> >>> 
> >>   A    B   x    y     z
> >> 1 1 10.0 100 1000 10000
> >> 2 2 19.8 200 2000 20000
> >> 3 3 19.8 300 3000 30000
> >> 4 4 19.8 400 4000 40000
> >> 5 5 30.0 500 5000 50000
> >> 
> >
> > [...]
> >
> > 
> >> b.na <- test$B
> >> b.na[c(F,diff(test$B)<1)] <- NA
> >> b.na
> >> 
> It is not clear that this works. What if test$B[4] is 20.9 instead of 
20.3?
> 
> There are cases that just don't vectorize. Brute force may be the only
> solution. It may be slow, but it is available.

That is what i meant with data in chunks. What if test$B looks like
seq(1, 10, .1).
I expected, maybe wrongly, that the data looks like

(n1...nn, big step1, m1...mn, big step2, ....)

and that David wants simply change this to NNNNNNN,MMMMM,....

And there are other ifs is B always increasing?

> 
> X <- test$B
> for (i in seq_along(X[-1]))
>    if (X[i] - X[i-1] < 1)
>       X[i] <- X[i-1]
> test$B <- X

X[i-1] is X[0] which does not work, you probably meant X[i+1]

Regards
Petr

> 
> -- 
>    O__  ---- Peter Dalgaard             Øster Farimagsgade 5, Entr.B
>   c/ /'_ --- Dept. of Biostatistics     PO Box 2099, 1014 Cph. K
>  (*) \(*) -- University of Copenhagen   Denmark          Ph:  (+45) 
35327918
> ~~~~~~~~~~ - (p.dalgaard at biostat.ku.dk)                  FAX: (+45) 
35327907
> 
>