Regex on assertr output

assertr_regex.R

library(dplyr)
library(assertr)

# if we have the command
data.frame(x = 1:2) %>% assert(in_set(1), x)
#> Column 'x' violates assertion 'in_set(1)' 1 time
#>   index value
#> 1     2     2
#> Error: assertr stopped execution

# this produces the output
out_txt <- "Column 'x' violates assertion 'in_set(1)' 1 time"

# the original suggestion to extract the column name was
gsub("^Column \'([a-zA-Z]+)\' .*$", '\\1', out_txt)
#> [1] "x"

# which works fine.

# if, however, the columns have real-world names, we have 
data.frame(c3po.x = 1:2) %>% assert(in_set(1), c3po.x)
#> Column 'c3po.x' violates assertion 'in_set(1)' 1 time
#>   index value
#> 1     2     2
#> Error: assertr stopped execution

# which produces the output
out_txt <- "Column 'c3po.x' violates assertion 'in_set(1)' 1 time"

# then the regex chokes and returns the whole string
gsub('^Column \'([a-zA-Z]+)\' .*$', '\\1', out_txt)
#> [1] "Column 'c3po.x' violates assertion 'in_set(1)' 1 time"

# a better regex might be (\\S == non-whitespace)
gsub('Column \'(\\S*?)\'.*$', '\\1', out_txt)
#> [1] "c3po.x"