[R] creating a new column assigning values of other columns

[R. Michael Weylandt]

It looks like part of your problem is that some of your time/date
variables are stored as factors rather than actual times / dates. Use
str() to see which ones and try to convert those. You'll need this
format string: format = "%d/%m/%y %H:%M:%S" for the ones that are
currently factors.

As regards getting POSIXct out of ifelse(), yes -- that seems to be an
infelicity, but I'm  sure its easily remedied. You just need:

as.POSIXct( ifelse( your_code_here) , origin = "1970-01-01")

Don't add a format string since that refers to the format of the input
(if you're giving a character input), not the output (which is
standardized in the definition of POSIXct)

Michael

On Sun, May 6, 2012 at 5:13 AM, Santiago Guallar <sguallar at yahoo.com> wrote:
> Hi Michael,
> Yes, I tried ifelse() before but this function returns a numeric value. When
> I try to convert it back to POSIXct I get a <NA>*. If I use if else I get a
> POSIXct output although it does not return a correct answer (it only returns
> y$timepos even when the condition "h<9" fails to be met).
> Attached the outputs you suggested (thanks, by the way, I didn't know
> dput()). Hope they go through this time.
>
> Thank you,
>
> Santi
>
> *niga$isnight<- as.POSIXct( niga$nit, tz="GMT", format="%Y-%m-%d %H:%M:%S",
> origin="2007-06-19" )
>
>
>
> From: R. Michael Weylandt <michael.weylandt at gmail.com>
> To: Santiago Guallar <sguallar at yahoo.com>
> Cc: "r-help at r-project.org" <r-help at r-project.org>
> Sent: Sunday, May 6, 2012 2:18 AM
> Subject: Re: [R] creating a new column assigning values of other columns
>
> Bahhhh -- far too much work to recreate (and I don't think you sent us
> the file "act.lig"): here's a much better route:
>
> Go to the step immediately before you're in trouble and use dput() on
> your data. R will print out a nice plaintext representation that we
> can copy and paste and reproduce *exactly* without having to do all
> that you show below.
>
> Incidentally, your warning message suggests you should be using
> ifelse() instead of if.
>
> To compare:
>
> x <- seq(-3, 3)
> abs.x.wrong <- if(x < 0) -x else x # Warning message gives some hint
> abs.x.right <- ifelse(x < 0, -x, x)
>
> Hope this helps,
>
> Michael
>
> On Sat, May 5, 2012 at 5:09 PM, Santiago Guallar <sguallar at yahoo.com> wrote:
>> Hello,
>>
>> I have to create a new column from the values of other columns of a data
>> frame. The new column (y$n) is created imposing a condition (using a third
>> variable y$h) that assigns the values of two time variables (y$b and
>> y$timepos). Here's the piece of code to get there (using the attached
>> files):
>>
>> xact <- read.table("act.lig", sep = ',',
>> col.names=c("ok","time","secs","act"))
>> xlig <- read.table("lig.txt", sep = ',',
>> col.names=c("ok","time","secs","lig"))
>> w<- merge(xact, xlig, by = c("time" ,"secs"), all = TRUE, sort=F)
>> require(reshape)
>> z <- cbind(w, colsplit(w$time, split=" ", names=c("date", "clock")))
>> zh<-cbind(z, colsplit(z$clock, split=":", names=c("h","m","s")))
>> zhd<- cbind(zh, colsplit(zh$date, split="/", names=c("d","mo","y")))
>> night <- subset(zhd, zh$lig<6 & zhd$h<9 | zh$lig<6 & zhd$h>21)
>> night$timepos<-as.POSIXct(night$time, tz="GMT", format="%d/%m/%y
>> %H:%M:%S")
>> a=night$timepos - as.difftime( 1, units="days" )
>> nighta<-cbind(night,a)
>> y<- cbind(nighta, b=as.character(a, tz= "GMT", format= "%Y-%m-%d"))
>> y$n<-with(y, if (h>=0 & h<9) {b} else {timepos}) ## Missing warnings
>>
>>                                                                             In
>> if (h >= 0 & h < 9) { :
>>
>>                                                                             condition
>> has length > 1 and only the first element will be used
>>
>> How can I go around this problem and get the new column?
>>
>> Thank you,
>>
>> Santi
>> ______________________________________________
>> R-help at r-project.org mailing list
>> https://stat.ethz.ch/mailman/listinfo/r-help
>> PLEASE do read the posting guide
>> http://www.R-project.org/posting-guide.html
>> and provide commented, minimal, self-contained, reproducible code.
>>
>
>
>
>