Remove a column from a data frame by name

Question

Possible Duplicate:
remove an entire column from a data.frame in R

Is there a better way to remove a column by name from a data frame than the following?

Orange[colnames(Orange) != "Age"]

I've tried the following and I get errors:

> Orange[-"Age"]
Error in -"age" : invalid argument to unary operator
> Orange[,-"Age"]
Error in -"age" : invalid argument to unary operator
> Orange[[,-"Age"]]
Error in -"age" : invalid argument to unary operator

score 6 · Answer 1 · answered Jul 19 '12 at 16:22

You can set the column to NULL

> dat <- data.frame(a = 1, b = 1, c = 1)
> dat
  a b c
1 1 1 1
> dat$a <- NULL
> dat
  b c
1 1 1
> dat["b"] <- NULL
> dat
  c
1 1

Someone will come along and point out that data.frame will makes lots of copies of the data to do this simple task. When data gets big (millions of rows), this will take a lot of time and may not work due to memory constraints. If that's going to be an issue, use data.table and the := operator:

library(data.table)
> dt <- data.table(a = 1, b = 1, c = 1)
> dt[,a:=NULL]
     b c
[1,] 1 1

Could you add your answer to the question I linked to, so quality answers aren't dispersed across multiple questions? — Joshua Ulrich, Jul 19 '12 at 16:44

score 3 · Answer 2 · answered Jul 19 '12 at 16:21

3

Try:

Orange[-match("age",names(Orange))]
   Tree circumference
1     1            30
2     1            58
3     1            87
4     1           115
...

answered Jul 19 '12 at 16:21

James

65,548
14
155
193

There is a much easier way, setting it to NULL, but it is discussed in the duplicate question (see above). – David Robinson Jul 19 '12 at 16:24

Remove a column from a data frame by name

2 Answers2

Linked