0

I'm trying to get agrep to match seconds to second and not millisecond, but there doesn't seem to be any value of costs to accomplish this.

I am especially confused that there's no value of the cost for deletions/insertions that seems to do the trick -- as I see it, second is one deletion from seconds whereas millisecond is one deletion and 5 insertions.

(warning that the lapply might take a while... you would get the same result for length.out = 10 and 0:10 much quicker)

rng = c(seq(0, 1, length.out = 20), 0:100)
x = expand.grid(insertions = rng, substitutions = rng, deletions = rng)

units = c("millisecond", "second", "minute", "hour", "day",
          "week", "month", "quarter", "year")
x$match = lapply(seq_len(nrow(x)), function(ii)
  agrep('second', units, value = TRUE, costs = x[ii, ]))

x$match_which = sapply(x$match, paste, collapse = '|')

sort(table(x$match_which))
#             millisecond|second|minute|hour|week|month|year 
#                                                         57 
#     millisecond|second|minute|hour|week|month|quarter|year 
#                                                      13276 
#                                   millisecond|second|month 
#                                                      23316 
#                    millisecond|second|minute|month|quarter 
#                                                      37842 
#                          millisecond|second|minute|quarter 
#                                                     251480 
# millisecond|second|minute|hour|day|week|month|quarter|year 
#                                                     409865 
#                                         millisecond|second 
#                                                    1035725

What am I missing here? Is there no way to accomplish my task (match seconds to second and not millisecond) with agrep?

MichaelChirico
  • 33,841
  • 14
  • 113
  • 198
  • I might be off, but does it help to do `agrep("^second", ...)` to make sure the matched pattern starts with `"second"`? – Lennyy Mar 27 '19 at 06:26
  • @Lennyy could be... but would need `fixed = FALSE` to do regex matching... still not sure why it wouldn't work with `fixed = TRUE` – MichaelChirico Mar 27 '19 at 08:40

0 Answers0