Learn R Programming

metatools (version 0.2.0)

build_from_derived: Build a dataset from derived

Description

This function builds a dataset out of the columns that just need to be pulled through. So any variable that has a derivation in the format of 'dataset.variable' will be pulled through to create the new dataset. When there are multiple datasets present, they will be joined by the shared key_seq variables. These columns are often called 'Predecessors' in ADaM, but this is not universal so that is optional to specify.

Usage

build_from_derived(
  metacore,
  ds_list,
  dataset_name = deprecated(),
  predecessor_only = TRUE,
  keep = FALSE
)

Value

dataset

Arguments

metacore

metacore object that contains the specifications for the dataset of interest.

ds_list

Named list of datasets that are needed to build the from. If the list is unnamed,then it will use the names of the objects.

dataset_name

[Deprecated] Optional string to specify the dataset that is being built. This is only needed if the metacore object provided hasn't already been subsetted.
Note: Deprecated in version 0.2.0. The dataset_name argument will be removed in a future release. Please use metacore::select_dataset to subset the metacore object to obtain metadata for a single dataset.

predecessor_only

By default TRUE, so only variables with the origin of 'Predecessor' will be used. If FALSE any derivation matching the dataset.variable will be used.

keep

String to determine which columns from the original datasets should be kept

  • "FALSE" (default): only columns that are also present in the ADaM specification are kept in the output.

  • "ALL": all original columns are carried through to the ADaM, including those that have been renamed. e.g. if DM.ARM is a predecessor to DM.TRT01P, both ARM and TRT01P will be present as columns in the ADaM output.

  • "PREREQUISITE": columns are retained if they are required for future derivations in the specification. Additional prerequisite columns are identified as columns that appear in the 'derivation' column of the metacore object in the format "DATASET.VARIABLE", but not as direct predecessors. Predecessors are defined as columns where the derivation is a 1:1 copy of a column in a source dataset. e.g. derivation = "VS.VSTESTCD" is a predecessor, while derivation = "Value of VS.VSSTRESN where VS.VSTESTCD == 'Heart Rate'" contains both VS.VSTESTCD and VS.VSSTRESN as prerequisites, and these columns will be kept through to the ADaM.

Examples

Run this code
library(metacore)
library(haven)
library(magrittr)
load(metacore_example("pilot_ADaM.rda"))
spec <- metacore %>% select_dataset("ADSL")
ds_list <- list(DM = read_xpt(metatools_example("dm.xpt")))
build_from_derived(spec, ds_list, predecessor_only = FALSE)

Run the code above in your browser using DataLab