Error in `filter()`: Input `..1` must be of size 4 or 1, not size 0-CodePudding

I'm very new to R. I have this data frame named "log"

  a b c d e 
x 1 2 3 4 5 
y 1 na TRUE 2 five
z 1 2 3 TRUE FALSE
t TRUE TRUE FALSE FALSE five

I run this code that goes well.

> filter(log, grepl("^[A-Za-z] $", log$b)==TRUE)

Output a b c d e y 1 na TRUE 2 five t TRUE TRUE FALSE FALSE five

If I try to put this code in a function where object = log (my data frame) and column = a column of my data frame :

wrong <- function(object, column){
      filter(object, grepl("^[A-Za-z] $", object$column)==TRUE)
}

I get this error

Error in `filter()`:
! Problem while computing `..1 = grepl("^[A-Za-z] $", object$column) == TRUE`.
✖ Input `..1` must be of size 4 or 1, not size 0.
---
Backtrace:
     ▆
  1. ├─global wrong(log, b)
  2. │ ├─dplyr::filter(...)
  3. │ └─dplyr:::filter.data.frame(object, grepl("^[A-Za-z] $", object$column) == TRUE)
  4. │   └─dplyr:::filter_rows(.data, ..., caller_env = caller_env())
  5. │     └─dplyr:::filter_eval(dots, mask = mask, error_call = error_call)
  6. │       ├─base::withCallingHandlers(...)
  7. │       └─mask$eval_all_filter(dots, env_filter)
  8. ├─dplyr:::dplyr_internal_error(...)
  9. │ └─rlang::abort(class = c(class, "dplyr:::internal_error"), dplyr_error_data = data)
 10. │   └─rlang:::signal_abort(cnd, .file)
 11. │     └─base::signalCondition(cnd)
 12. └─dplyr (local) `<fn>`(`<dpl:::__>`)
 13.   └─rlang::abort(bullets, call = error_call, parent = skip_internal_condition(e))

I don't understand what's wrong, why the first code works but the second dosen't work. I tried to modify my code but it didn't work. Any help appreciated

Thanks!

CodePudding user response：

The $column will not work as column it will look for the literal column instead of the value. Also, not clear if we are passing unquoted or quoted i.e. string column name in the function. If we are passing unquoted, use {{}} operator. Also, note that grepl returns a logical vector as output i.e. TRUE/FALSE, therefore, there is no need to do another comparison (== TRUE) on top of it

library(dplyr)
f1 <- function(object, column){
       filter(object, grepl("^[A-Za-z] $", {{column}}))
 }

-testing

> f1(log, b)
     a    b     c     d    e
y    1   na  TRUE     2 five
t TRUE TRUE FALSE FALSE five

It may not work if we est as f1(log, "b"). To make it more general, i.e. to pass either unquoted or quoted, convert to symbol with ensym and evaluate (!!)

f2 <- function(object, column)
    {
    filter(object, grepl("^[A-Za-z] $", !! rlang::ensym(column)))
   }

-testing

> f2(log, b)
     a    b     c     d    e
y    1   na  TRUE     2 five
t TRUE TRUE FALSE FALSE five
> f2(log, "b")
     a    b     c     d    e
y    1   na  TRUE     2 five
t TRUE TRUE FALSE FALSE five

Another option is also to make use of .data[[column]] if we pass as string or if there are one or more columns, use if_any/if_all as well

f3 <- function(object, column){
      filter(object, if_any(all_of(column), ~ grepl("^[A-Za-z] $", .x)))
}

-testing

> f3(log, "b")
    a    b     c     d    e
y    1   na  TRUE     2 five
t TRUE TRUE FALSE FALSE five
> f3(log,  c("b", "c", "e"))
     a    b     c     d     e
y    1   na  TRUE     2  five
z    1    2     3  TRUE FALSE
t TRUE TRUE FALSE FALSE  five

data

log <- structure(list(a = c("1", "1", "1", "TRUE"), b = c("2", "na", 
"2", "TRUE"), c = c("3", "TRUE", "3", "FALSE"), d = c("4", "2", 
"TRUE", "FALSE"), e = c("5", "five", "FALSE", "five")), 
class = "data.frame", row.names = c("x", 
"y", "z", "t"))