2

Is there possibility to write function f(), which I can use for subsetting columns set which we can use strictly like this : data[, f(c("var1", "var2", "var3"))] without using data as argument,f() could be something like : 

f <- function(data_frame, minus_colnames)
     return colnames(data_frame)[which(!colnames(data_frame)%in%minus_colnames)]

but can we somehow call f without explicitly pointing to data dataframe like :

data[, f(c("var1", "var2", "var3"))]

not like

data[, f(data, c("var1", "var2", "var3"))] (and without dplyr) ?

That would be most suitable syntax for me

Qbik
  • 5,885
  • 14
  • 62
  • 93

1 Answers1

4
subset(data, select= -c(var1, var2, var3))

or

data[,!names(data) %in% c("var1","var2","var3"))]

or as @DavidArenburg suggests

setdiff(names(data), c("var1", "var2", "var3")

dplyr::select is another option; I'm sure there are data.table solutions too.

Ben Bolker
  • 211,554
  • 25
  • 370
  • 453
  • The dplyr option could be `select(data, -num_range("var", 1:3))`, for example. (+1) – talat Nov 11 '15 at 13:21
  • To finish it off `dt[, !paste0('var', 1:3), with=F]` and `dt[, paste0('var', 1:3) := NULL]` for data.table – Pierre L Nov 11 '15 at 13:31