grepl often faster with perl = T

[KyleHaynes]

I notice this library has quite a few grepl calls. Not sure if you're aware but using perl is often considerably faster ...

example

> vec <- c(" ", "    ", "asdas", "dsdsd", letters, LETTERS)
> vec <- sample(vec, 10000000, replace = T)
> system.time(grepl("^\\s*$", vec))
   user  system elapsed 
   1.38    0.00    1.39 
> system.time(grepl("^\\s*$", vec, perl = T))
   user  system elapsed 
   0.72    0.03    0.75 
> system.time(grepl("^\\s*$", vec))
   user  system elapsed 
   1.32    0.00    1.33 
> system.time(grepl("^\\s*$", vec, perl = T))
   user  system elapsed 
   0.69    0.00    0.69 

[trinker]

Hmmm. Very nice. Would you be willing to (1) test this on longer strings as well? And (2) make a pr with changes after checking the speed on each instance to make sure there is an improvement?

[KyleHaynes]

1. yep, 2) yep. Will get to this in the next few days

[trinker]

@KyleHaynes checking to see if you had done any work with this

[KyleHaynes]

@trinker extremely sorry, been completely flat chat with high priority work. I won't get to this for about a week.

[trinker]

Related: https://stackoverflow.com/a/47251004/1000343