Skip to contents

The input dataset is split into slices (subsets) and for each slice the derivation is called separately. Some or all arguments of the derivation may vary depending on the slice.

Usage

slice_derivation(dataset, derivation, args = NULL, ...)

Arguments

dataset

Input dataset

derivation

Derivation

args

Arguments of the derivation

A param() object is expected.

...

A derivation_slice() object is expected

Each slice defines a subset of the input dataset and some of the parameters for the derivation. The derivation is called on the subset with the parameters specified by the args parameter and the args field of the derivation_slice() object. If a parameter is specified for both, the value in derivation_slice() overwrites the one in args.

Value

The input dataset with the variables derived by the derivation added

Details

For each slice the derivation is called on the subset defined by the filter field of the derivation_slice() object and with the parameters specified by the args parameter and the args field of the derivation_slice() object. If a parameter is specified for both, the value in derivation_slice() overwrites the one in args.

  • Observations that match with more than one slice are only considered for the first matching slice.

  • Observations with no match to any of the slices are included in the output dataset but the derivation is not called for them.

Author

Stefan Bundfuss

Examples

library(tibble)
library(stringr)
advs <- tribble(
  ~USUBJID, ~VSDTC,       ~VSTPT,
  "1",      "2020-04-16", NA_character_,
  "1",      "2020-04-16", "BEFORE TREATMENT"
)

# For the second slice filter is set to TRUE. Thus derive_vars_dtm is called
# with time_imputation = "last" for all observations which do not match for the
# first slice.
slice_derivation(
  advs,
  derivation = derive_vars_dtm,
  args = params(
    dtc = VSDTC,
    new_vars_prefix = "A"
  ),
  derivation_slice(
    filter = str_detect(VSTPT, "PRE|BEFORE"),
    args = params(time_imputation = "first")
  ),
  derivation_slice(
    filter = TRUE,
    args = params(time_imputation = "last")
  )
)
#> # A tibble: 2 x 5
#>   USUBJID VSDTC      VSTPT            ADTM                ATMF 
#>   <chr>   <chr>      <chr>            <dttm>              <chr>
#> 1 1       2020-04-16 NA               2020-04-16 23:59:59 H    
#> 2 1       2020-04-16 BEFORE TREATMENT 2020-04-16 00:00:00 H