Operator "[<-" in RStudio and R -

by accident i've encountered strange behaviour of "[<-" operator. behaves differently depending on order of calls , whether i'm using rstudio or ordinary rgui. make myself clear example.

x <- 1:10 "[<-"(x, 1, 111) x[5] <- 123 

as far know, first assigment shouldn't change x (or maybe i'm wrong?), while second should do. , in fact result of above operations is

x [1]  1  2  3  4  123  6  7  8  9 10 

however, when perform these operations in different order, results different , x has changed! meaningly:

x <- 1:10 x[5] <- 123 "[<-"(x, 1, 111) x [1] 111   2   3   4   123   6   7   8   9  10 

but happens when i'm using plain r! in rstudio behaviour same in both options. i've checked on 2 machines (one fedora 1 win7) , situation looks same. know 'functional' version ("[<-"(x..)) never used i'm curious why happening. explain that?

==========================

edit: ok, comments reason x <- 1:10 has type 'integer' , after replacing x[5] <- 123 it's 'double'. still remains question why behaviour different in rstudio? restart r session , doesn't change anything.

rstudio's behavior

rstudio's object browser modifies objects examines in way forces copying upon modification. specifically, object browser employs @ least 1 r function call internally forces evaluation of object, in process resetting value of object's named field 1 2. r-internals manual:

when object altered, named field consulted. value of 2 means object must duplicated before being changed. [...] value of 1 used situations [...] in principle 2 copies of exist duration of computation [...] no longer, , primitive functions can optimized avoid copy in case.

to see object browser modifies named field ([nam()] in next code block), compare results of running following lines. in first, both 'lines' run together, rstudio has no time 'touch' x before structure queried. in second, each line pasted in separately, x modified before examined.

## pasted in x <- 1:10; .internal(inspect(x)) # @46b47b8 13 intsxp g0c4 [nam(1)] (len=10, tl=0) 1,2,3,4,5,...  ## pasted in delay between lines x <- 1:10 .internal(inspect(x)) # @42111b8 13 intsxp g0c4 [nam(2)] (len=10, tl=0) 1,2,3,4,5,...  

once named field set 2, [<-(x, ...) not modify original object. pasting following rstudio @ once modifies x, while pasting in line-by-line not:

x <- 1:10 "[<-"(x, 1, 111) 

one more consequence of rstudio's object browser makes operations slower otherwise be. again, compare same 2 commands first pasted in together, , 1 @ time:

## pasted in x <- 1:5e7 system.time(x[1] <- 9l) #    user  system elapsed  #       0       0       0   ## pasted in 1 @ time x <- 1:5e7 system.time(x[1] <- 9l) #    user  system elapsed  #    0.11    0.04    0.16  

---

variable behavior of [<- in r

the behavior of [<- w.r.t. modifying vector x depends on storage types of x , of element being assigned it. explains r's behavior not rstudio's.

in r, when [<- either appends vector x, or performs subassignment requires x's type modified, x copied , value returned not overwrite pre-existing variable x. (to need x <- "[<-(x, 2, 100).

so, neither of following modify x

x <- 1:2         ## note: typeof(x) --> "integer"  ## subassignment requires x coerced "numeric" type "[<-"(x, 2, 100) ## note: typeof(100) --> "numeric" x  # [1]   1   2  ## appending x "[<-"(x, 3, 100l) x # [1]   1   2 

whenever possible, though, r allow [<- function modify x directly reference (i.e. without making copy). "possible" here includes cases in sub-assignment doesn't require x's type modified.

so of following modify x

x <- c(0i, 0i, 0i, 0i) "[<-"(x, 1, true) "[<-"(x, 2, 20l) "[<-"(x, 3, 3.14) "[<-"(x, 4, 5+5i) x # [1]  1.00+0i 20.00+0i  3.14+0i  5.00+5i