This refactors and speeds up guess_dates. Now, instead of taking 2.5 seconds for 5K messy date entries, it takes 1.25 seconds. It's still not phenomenally fast, but it's faster and that's what matters.
This also changes the excel argument from a logical to a character, but the default will work the same.
The old version was slow:
remotes::install_github("reconhub/linelist@0e6ac439963714ee0ca17a99237f554f6c183b6d")
library(linelist)
md <- messy_data(10000)$"messy/dates"
system.time(guess_dates(md, err = 1))
#> user system elapsed
#> 3.530 0.076 3.605
By contrast, the new version doubles the speed.
remotes::install_github("reconhub/linelist@00f0e3db9d4264aea4bdccc2bccb85c9b4e72227")
library(linelist)
md <- messy_data(10000)$"messy/dates"
system.time(guess_dates(md, err = 1))
#> user system elapsed
#> 1.714 0.092 1.806
This refactors and speeds up guess_dates. Now, instead of taking 2.5 seconds for 5K messy date entries, it takes 1.25 seconds. It's still not phenomenally fast, but it's faster and that's what matters.
This also changes the excel argument from a logical to a character, but the default will work the same.
The old version was slow:
By contrast, the new version doubles the speed.
Created on 2019-05-15 by the reprex package (v0.2.1)
It now also has arguments for excel dates that include old windows (which addresses #73).