Hi everyone,
I try to add many vectors (L1,L2,L3....) to many list objects (a.list,
b.list....) in a workspace. Somethings like below, but it is not working. Any
suggestions will be appreciated. Best, John
lf=ls(pattern=".lst")
for (x in listfiles) {
dat=read.delim(x,header=F)
for (i in 1: lf) {
assign(i$add,as.numeric(dat[,3]))
#or i$add=as.numeric(dat[,3]
names(i)[names(i)=="add"]=substr(x,1,5)
print (i[1:3,])
}
}
[[alternative HTML version deleted]]
On Aug 1, 2012, at 8:11 AM, John linux-user wrote:> Hi everyone, > > I try to add many vectors (L1,L2,L3....) to many list objects > (a.list, b.list....) in a workspace. Somethings like below, but it > is not working. Any suggestions will be appreciated. Best, John > > lf=ls(pattern=".lst") > > for (x in listfiles) { > dat=read.delim(x,header=F)Presumably that would fail since 'listfiles' has not been defined. did you mean 'lf'? If you did,then wouldn't the second line overwrite all the early values of "dat" leaving only the last one? Perhaps: datfils <- list() for (x in listfiles) { datfils[x] <- read.delim(x,header=F)> > for (i in 1: lf) {And that would fail because 'lf' is a character vector, and it's not meaningful to specify such a range. Try instead: for (i in names(datfils[x]) ) { # # which will then iterate over the names of the files which are now also the names of the list elements> assign(i$add,as.numeric(dat[,3]))But since 'i' is a length-1 character vector, the expression `i$add` will be meaningless. The "$" operator does not do function calling in R unless you do fancy things with environments, and you cannot "sub- assign" in that manner, at least not with the assign() function. Try instead: assign(i, as.numeric(datfils[x][,3])) names(i)[length(i)] <- "add" Or: i <- transform(i, add=datfils[x][,3] )> #or i$add=as.numeric(dat[,3] > names(i)[names(i)=="add"]=substr(x,1,5)I'm not sure these would be doing the same thing. What was your goal here?> > print (i[1:3,]) > } > }David Winsemius, MD Alameda, CA, USA
On Aug 3, 2012, at 2:11 PM, R. Michael Weylandt wrote:> On Fri, Aug 3, 2012 at 2:53 PM, John linux-user <johnlinuxuser at yahoo.com > > wrote: >> Maybe what I previously posted was not clear enough or something >> else. All vectors L1,L2.. and objects (e.g. a.list, b.list, c.list) >> already exit or easily to be created in a workspace. Do not worry >> about those. The key question is how to systematically append/ >> assign these vectors to many objects (hundreds) in a loop without >> explicitly typing these object names one by one. Please refer to my >> code posted below to detail my question. I really appreciate your >> ideas/suggestions. John >> > > Like David, I find this still somewhat confusing, but perhaps this > will get you started: > > a3 <- 1:4 > b3 <- 6 > c3 <- letters > > a_new_thing <- do.call(list, lapply(ls(pattern = "3"), get)) > > old_thing <- list("cow") > > a_longer_old_thing <- c(old_thing, a_new_thing) > > adapt as needed.Heh. Users of this should realize that Michael probably did this in clean session. If you run it with any other items in the workspace that have "3" in their names they get wrapped up too. I had a 300 line dataframe as well as a function whose values both got gathered up. In the case where one needed to isolate a disjoint group of items, one might want to create an environment in which to segregate those items. the ls function can be directed to look in only one environment. I wondered if this were the route intended: for( i in seq_along(ls(patt=".lst") ) ) { assign( ls(patt=".lst")[i], # the name of the i-th ".lst"- item append( sapply(ls(patt=".lst$"), get)[i], # the original value sapply(ls(patt="^L\\d"), get)[i] ) )} # the appended i-th "L"-value I think it is far from optimal to have a loose collection of objects in the workspace. Much better to have them all in one list or an environment. -- David Winsemius, MD Alameda, CA, USA