Add a content-hash series id to a data.table

Adds time_series_label, the identity values joined by sep, and time_series_id, the xxhash64 digest of that label. The id depends only on the identity values, so a series gets the same id in every table.

Usage

set_time_series_id(d, id_cols, sep = "\037")

Arguments

d

A data.table. The function changes it by reference.

id_cols

The identity columns that define one series.

sep

The separator in the label. The default is the ASCII unit separator, "\037".

Value

d, invisibly.

Details

csfmt_ensemble_v3() and csfmt_reporting_triangle_v3() call this function for you.

See also

vignette("pipeline", package = "csalert"), which shows the ensemble format.

Other ensemble format functions: csfmt_ensemble_v3(), print.csfmt_ensemble_v3(), validate_ensemble()

Examples

d <- data.table::data.table(
  location_code = c("nation", "nation", "county03"),
  age = "total",
  isoyearweek = c("2023-01", "2023-02", "2023-01"),
  numerator = c(10, 12, 4)
)
set_time_series_id(d, id_cols = c("location_code", "age"))

# the two nation rows share one id, and the county row has its own
d[]
#>    location_code    age isoyearweek numerator   time_series_id
#>           <char> <char>      <char>     <num>           <char>
#> 1:        nation  total     2023-01        10 0ec581372be7c933
#> 2:        nation  total     2023-02        12 0ec581372be7c933
#> 3:      county03  total     2023-01         4 0d06fc1a57d1a2a5
#>    time_series_label
#>               <char>
#> 1:   nation\037total
#> 2:   nation\037total
#> 3: county03\037total