In R what's the difference between [[X]] and [, X] when selecting vectors

Question

library(tidyverse)
df0 <- data.frame(col1 = c(5, 2), col2 = c(6, 4))
df1 <- data.frame(col1 = c(5, 2), 
                  col2 = c(6, 4),
                  col3 = ifelse(apply(df0[, 1:2], 1, sum) > 10 & 
                                  df0[, 2] > 5, 
                                "True",
                                "False"))
df2 <- as_tibble(df1)

I've got my data frame df1 above. I've basically "copied" it as a tibble df2 . Let's mimic an analysis for this df1 data frame and df2 tibble.

identical(df1[[2]], df1[, 2])
# [1] TRUE
identical(df2[[2]], df2[, 2])
# [1] FALSE

Since df1 and df2 are essentially the "same", why do I get the TRUE/FALSE dichotomy in my code block above. What is the tibble() property that has changed?

The same question asked another way - what is the difference between [[X]] and [, X] , when applied to base R, and also when used in the tidyverse?

Answer 1

Since all lists are vectors, we can think of this in terms of list subsetting. Take for instance:

L <- list(A = c(1, 2), B = c(1, 4))
L[[2]]

This Extract s the second element of the list. Extrapolate this to:

df1[[2]] 

We get the same output as df1[, 2] hence identical(df1[[2]], df1[, 2]) returns TRUE . The second part is to do with tibble structure ie:

typeof(as_tibble(df1)[[2]])
[1] "double"
typeof(as_tibble(df1[, 2]))
[1] "list"

The second is a list while the first is a vector hence identical returns FALSE .

Objects of class tbl_df have: (From the docs)

A class attribute of c("tbl_df", "tbl", "data.frame") .

A base type of "list", where each element of the list has the same NROW().

A names attribute that is a character vector the same length as the underlying list.

A row.names attribute, included for compatibility with the base data.frame class. This attribute is only consulted to query the number of rows, any row names that might be stored there are ignored by most tibble methods.