'Error in !`*tmp*` : invalid argument type' error when using !! operator in R-CodePudding

I am trying to programmatically initialize some variables in R so that the variable name would be the evaluated content of the string.

Just this code:

library(dplyr)

v <- 'sum.of.ranfx'
new_v = sym(v)

!!new_v <- vector(mode = "list", length = 122)

fails with

Error in !`*tmp*` : invalid argument type

Google gives me no hits for this exact error. Here is an example accepted and upvoted SO answer whose syntax example I think I am following. Can you tell me what I'm doing wrong?

CodePudding user response：

You may use assign -

v <- 'sum.of.ranfx'
assign(v, vector(mode = "list", length = 122))

CodePudding user response：

raw !! is not accepted, so you should put ` around the !!new to make it work

`!!new`= vector(mode = "list", length = 122)

I think it's because ! is used to 'reverse' a T or a F

ex:

> !TRUE
[1] FALSE
> !!TRUE
[1] TRUE

So when you try to create a variable without specifying the ! should be used as a character, R tries to perform this reverse operation. Was this understandable?

CodePudding user response：

We may use list2env

v <- 'sum.of.ranfx'
list2env(setNames(list(vector(mode = "list", length = 122)), v), .GlobalEnv)

-checking the object

> head(sum.of.ranfx)
[[1]]
NULL

[[2]]
NULL

[[3]]
NULL

[[4]]
NULL

[[5]]
NULL

[[6]]
NULL

CodePudding user response：

For curious future readers, here is what I was doing wrong: attaching the wrong package. I needed library(rlang) in addition to library(dplyr), and I was confused by the fact that one piece of example code on the internet was explicitly showing dplyr being attached, and not rlang, but that's not a feature of the SO answer I linked in the question. When rlang is attached it runs without error.