I am trying to understand a problem regarding the pull() function using these two examples:

df <- data.frame(x = c("ex", "ex"), y = c("why", "why"))
d <- df
x <- "y"

pull(d, x)

d <- select(df, -x)

pull(d, x)
pull(d, y)

# - - - - - - 

Why does this output look like the pull() function is working well but the output before was not working?

library(dplyr)

# Using the dataframe “mtcars”
df <- mtcars

# Define a variable
x <- 'mpg'

# Using pull() to extract column "mpg"
resultado <- pull(df, x)
print(resultado)

# Remove column "mpg"
df <- select(df, -x)

# Using pull() again with variable x, which still contains "mpg"
resultado <- pull(df, x)
print(resultado)

In this second case thepull(df, x) should be different right? Not an error.

In both chunks of code I should have the same result which is a vector. But at the second code using mtcars as datframe I get this error:

Error in pull(): ! Can't extract columns that don't exist. ✖ Column mpg doesn't exist.

Considering this article: https://graphdr.github.io/data-stories/posts/2024-08-01-subtle-flaw-pull/