Why data frame column names different using = and <- in R

Question

Can any body explain why the below two data frames df1 and df2 are differing in their column names

  df1 <- data.frame(a = 1:5, b = 11:15)
  df1
  #   a  b
  # 1 1 11
  # 2 2 12
  # 3 3 13
  # 4 4 14
  # 5 5 15

  df2 <- data.frame(a <- 1:5, b <- 11:15)
  df2
  #   a....1.5 b....11.15
  # 1        1         11
  # 2        2         12
  # 3        3         13
  # 4        4         14
  # 5        5         15

In the first case you're passing named `...` arguments. In the second you are passing the result of a function and R `deparse`s the expression to name the column. The second case is similar to a call like `data.frame(1:5 + sum(3^2 - 10:3) - 28.8)`. R trying to make valind "names" uses something like `make.names(deparse(substitute(a <- 1:5)))`. Also, in the second case you've assigned a `a` and `b` object in your `.GlobalEnv` — alexis_laz, Sep 12 '16 at 10:52

Sandipan Dey · Answer 1 · 2016-09-12T12:01:40.350

If you want to have the column names as a, b, the correct syntax should be

data.frame(a=1:5, b=1:5)

The statement

data.frame(a <- 1:5, b <- 1:5)

R interprets it as if no column names are provided, so it treats the entire expression 'a <- 1:5' as the first column name, but there are 2 spaces and 3 illegal characters '<', '-', ':' that are not allowed in a column name, so each of them is changed to the character '.', hence you get the entire 'a....1.5' as the first column name, same goes for the second column.

score -1 · Accepted Answer · answered Sep 12 '16 at 10:52

The <- operator not only assigns objects but creates them in the parent environment unlike = operator

Renaming columns slightly:

df1 <- data.frame(a1 = 1:5, b1 = 11:15)



df1
#   a1  b1
# 1 1 11
# 2 2 12
# 3 3 13
# 4 4 14
# 5 5 15


#The objects are only created in the dataframee but not in environment  
# > exists(x = "a1")
#[1] FALSE
#> exists(x = "b1")
#[1] FALSE


#The objects are not only in created in the dataframe as well as in the environment

df2 <- data.frame(a2 <- 1:5, b2 <- 11:15)
df2
#   a2....1.5 b2....11.15
# 1        1         11
# 2        2         12
# 3        3         13
# 4        4         14
# 5        5         15

# > exists(x = "a2")
#[1] TRUE
#> exists(x = "b2")
#[1] TRUE

This is useful, but it is already provided in the other question I referenced in the comment. It would be more informative if you included an explanation about how `data.frame` treats the call to get column names. — jakub, Sep 12 '16 at 11:04

Why data frame column names different using = and <- in R

2 Answers2