Error: target of assignment expands to non-language object - r

Question

I am trying to dynamically clean up some column names for a large number of tables and I get the above error.
I have a gut feeling that I should be using quo but I have no idea on how to do that.
Any ideas?
The apply_alias applies a set of business rules to clean the names.
apply_alias <- function(l){
which(l=="Geography")
l[which(l=="Geography")] <- "GEO"
toupper(l)
}
The cleanup_column_names_tbl applies the alias_function to a list of table
cleanup_column_names_tbl <- function(PID){
for(p in PID){
names(get(paste0("tbl_",p))) <- apply_alias(names(get(paste0("tbl_",p))))
}
}
cleanup_column_names_tbl("14100287")
When I try to run it i get the following error message:
> cleanup_column_names_tbl("14100287")
Error in names(get(paste0("tbl_", p))) <- apply_alias(names(get(paste0("tbl_", :
target of assignment expands to non-language object
Sample data:
> dput(tbl_14100287[1,])
structure(list(V1 = 0L, REF_DATE = "1976-01", GEO = "Canada",
DGUID = "2016A000011124", `Labour force characteristics` = "Population",
Sex = "Both sexes", `Age group` = "15 years and over", Statistics = "Estimate",
`Data type` = "Seasonally adjusted", UOM = "Persons", UOM_ID = 249L,
SCALAR_FACTOR = "thousands", SCALAR_ID = 3L, VECTOR = "v2062809",
COORDINATE = "1.1.1.1.1.1", VALUE = 16852.4, STATUS = "",
SYMBOL = NA, TERMINATED = NA, DECIMALS = 1L), class = c("data.table",
"data.frame"), row.names = c(NA, -1L), .internal.selfref = <pointer: 0x000002123cf21ef0>)

Answer 1

You cannot assign a value to get since there is no function get<-. The right way of doing it would be something like the following.
apply_alias <- function(l){
l[which(l == "Geography")] <- "GEO"
toupper(l)
}
cleanup_column_names_tbl <- function(PID, envir = .GlobalEnv){
pid_full <- paste0("tbl_", PID)
res <- lapply(pid_full, function(p){
nms <- apply_alias(names(get(p)))
DF <- get(p)
names(DF) <- nms
DF
})
names(res) <- pid_full
list2env(res, envir = envir)
invisible(NULL)
}
cleanup_column_names_tbl("14100287")
names(tbl_14100287)
# [1] "V1" "REF_DATE"
# [3] "GEO" "DGUID"
# [5] "LABOUR FORCE CHARACTERISTICS" "SEX"
# [7] "AGE GROUP" "STATISTICS"
# [9] "DATA TYPE" "UOM"
#[11] "UOM_ID" "SCALAR_FACTOR"
#[13] "SCALAR_ID" "VECTOR"
#[15] "COORDINATE" "VALUE"
#[17] "STATUS" "SYMBOL"
#[19] "TERMINATED" "DECIMALS"

Answer 2

My solution:
Creating and expression and evaluating it. It is short but somehow does not feel like it is the proper way of doing things since I am breaking away from the functional paradigm of R.
cleanup_column_names_tbl <- function(PID){
for(p in PID){
expr1 <- paste0("names(", paste0("tbl_",p), ") <- apply_alias(names(", paste0("tbl_",p),"))")
eval(rlang::parse_expr(expr1))
}
}
Edit:
A slightly different way that:
avoids using strings
a little more flexible since it allows the use of strings and symbol
library(rlang)
test_df <- data.frame(a=1:10,b=1:10)
test_df2 <- data.frame(a=1:10,b=1:10)
fix_names <- function(df){
x <- ensym(df)
expr1 <- expr(names(!!x) <- toupper(names(!!x)))
eval(expr1, envir = parent.env(environment()))
# expr1
}
fix_names(test_df)
fix_names("test_df2")
names(test_df)
#> [1] "A" "B"
names(test_df2)
#> [1] "A" "B"