optimise numeric_dates

[SymbolixAU]

As per #59 , move the date conversion from data.frames out of the 'by-row' iteration and up to the top level.

[SymbolixAU]

Before

N <- 1e3
data <- data.frame(
  longitude = runif(N, -180, 180),
  latitude = runif(N, -180, 180),
  tooltip = sample(LETTERS, N, replace = TRUE),
  date = Sys.Date()
  #date_txt = as.character(Sys.Date())
)

microbenchmark::microbenchmark(
  fact = {
    js_str <- jsonify::to_json(data, numeric_dates = TRUE )
  },
  str = {
    js_fac <- jsonify::to_json(data, numeric_dates = FALSE )
  },
  times = 5
)

# Unit: milliseconds
# expr        min         lq       mean     median         uq        max neval
# fact   1.663784   1.731362   1.914341   1.802852   1.940997   2.432712     5
#  str 910.115947 927.891171 942.588014 929.894447 948.521252 996.517253     5

After

N <- 1e3
data <- data.frame(
  longitude = runif(N, -180, 180),
  latitude = runif(N, -180, 180),
  tooltip = sample(LETTERS, N, replace = TRUE),
  date = Sys.Date()
  #date_txt = as.character(Sys.Date())
)

microbenchmark::microbenchmark(
  fact = {
    js_str <- jsonify::to_json(data, numeric_dates = TRUE )
  },
  str = {
    js_fac <- jsonify::to_json(data, numeric_dates = FALSE )
  },
  times = 5
)

# Unit: milliseconds
# expr      min       lq     mean   median       uq      max neval
# fact 1.640394 1.709926 1.755007 1.758301 1.799105 1.867309     5
# str  2.459267 2.488134 2.808100 2.490519 2.693397 3.909183     5