Dates

[SymbolixAU]

Dates treated as numeric or character?

Numeric : more speed
Character : human-readable

jsonify::to_json( data.frame(dte = as.Date("2018-01-01") ) )
# [1] "[{\"dte\":17532.0}]"

jsonify::to_json( jsonify:::handle_dates( data.frame(dte = as.Date("2018-01-01") ) ) )
# [1] "[{\"dte\":\"2018-01-01\"}]"

---

n <- 1e6
df <- data.frame(id = 1:n, dte = sample(seq(as.Date("2018-01-01"), as.Date("2018-01-31"), length.out = n)))

library(microbenchmark)
microbenchmark(
  numeric = {
    jsonify::to_json( df )
  },
  character ={
    jsonify::to_json( jsonify:::handle_dates( df ) )
  },
  times = 5
)
# Unit: milliseconds
#      expr       min       lq     mean   median       uq      max neval
#   numeric  982.7153 1113.085 1117.633 1140.885 1149.160 1202.320     5
# character 4802.2583 4829.550 5130.051 4916.636 5146.547 5955.265     5

---

Decision

• always numeric
• always character
• date_as_numeric = TRUE argument in to_json(). If FALSE, use handle_dates()

[SymbolixAU]

I'm leaning towards the quickest option as the default, so leaving Dates as numeric, but providing the dates_as_numeric (or similarly named) argument.

[SymbolixAU]

Given this issue & correction , need a robust solution for lists too.

[SymbolixAU]

I'm not going to support numeric_dates in lists for v0.1.0.