Tabs are not trimmed when trim_ws = TRUE

[mirkhosro]

When using read_csv to read a CSV file that contains tabs on its rows, the trim_ws argument does not seem to have any effect. The tab characters do not get stripped anyway.
Take this CSV file for example

X,Y
x1,y1
x2	,y2
x3,y3	
x4,y4

There is a tab after x2 and after y3 (not visible). And the code below demonstrates the issue

> data <- read_csv("test.csv", trim_ws = TRUE, col_types = "cc")
> data[, 1]
# A tibble: 4 x 1
       X
   <chr>
1     x1
2 "x2\t"
3     x3
4     x4
> data[, 2]
# A tibble: 4 x 1
       Y
   <chr>
1     y1
2     y2
3 "y3\t"
4     y4

As you can see the tab characters are there and not trimmed.
I'm running the code using R 3.3.3 and readr 1.1.1. on macOS High Sierra.

[batpigandme]

Could you please turn this into a reprex so we can be sure we're troubleshooting the exact same thing? Thanks

[mirkhosro]

Here it is. I hope I've done it correctly.

library(readr)
csv_file <- "X,Y\nx1\t,y1\nx2,y2\t\nx3 ,y3 \n"
data <- read_csv(csv_file, trim_ws = TRUE, col_types = "cc")
data[, 1]
#> # A tibble: 3 x 1
#>        X
#>    <chr>
#> 1 "x1\t"
#> 2     x2
#> 3     x3
data[, 2]
#> # A tibble: 3 x 1
#>        Y
#>    <chr>
#> 1     y1
#> 2 "y2\t"
#> 3     y3

[mirkhosro]

(bump) this bug seems to be still there and left open, are you able to reproduce it using the reprex above?

[mirkhosro]

Thanks for the fix @jimhester . But it seems to only handle the space and tab characters and not "whitespace" in general. Wouldn't it better to capture white space using \s regex?

[jimhester]

No.

[mirkhosro]

Thanks for your extensive explanation :)