Skip to contents

identify/categorize columns from adsl

Source: R/adam_spec_adsl.R
adsl_identify.Rd

family of helper functions to identify columns to drop from adsl data set

Usage

adsl_identify(
  data,
  dict = NULL,
  type = c("dttm", "constant", "combined", "flag", "factor", "redundant"),
  dict_label = "label",
  dict_param = "param",
  id = "SUBJID",
  trt = "TRT01A",
  black_list = c("RANDNO", "SITEID", "SITENAM", "INVID", "INVNAM")
)

adsl_identify_dttm(data)

adsl_identify_constant(data)

adsl_identify_combined(
  data,
  dict = NULL,
  dict_label = "label",
  dict_param = "param"
)

adsl_identify_redundant(data, id, trt, clmn_flag)

adsl_identify_flag(data, dict, dict_param = "param", dict_label = "label")

adsl_identify_factor(
  data,
  id,
  clmn_flag = NULL,
  dict,
  dict_param = "param",
  dict_label = "label"
)

adsl_identify_factor_data(data)

Arguments

data

adsl-like data set in which to identify particular columns of interest

dict, dict_param, dict_label

dict is tibble as created by adsl_dict() where dict_param and dict_label indicate the columns in dict containing for parameter names (column names of data) and labels, resp.

type

character vector determining the categories of column types to identify. defaults all possible categories: dttm, constant, combined, flag, factor, redundant

id, trt

user-selected column names in data for ID and treatment column, defaulting to SUBJID and TRT01A, resp.

black_list

character vector of columns that should be dropped for most analyses, see details.

clmn_flag

(factor and redundants only) character vector of names identified as flags

Value

list with two top level entries, where to_remove is a list of column names from data that were identified as candidates for a given category and lev_list a list required to set factor level orders.

Details

Columns meeting the following criteria are returned

adsl_identify_dttm(): methods::is(.x, "Date") is TRUE, the label contains strings 'year', 'month', 'day', 'date' or 'time' (not case sensitive), class is one of 'difftime', 'hms', 'Period', 'POSIXct', 'POSIXt', 'Date'

adsl_identify_constant(): identification via janitor::remove_empty(which = 'cols'), janitor::remove_constant(na.rm = TRUE)

adsl_identify_redundant(): redundant columns to selected trt and id columns

adsl_identify_combined(): if labels (from dict) contain '/' and all parts are column names themselves

By default, black_list contains RANDNO, SITEID, SITENAM, INVID, INVNAM.

Authors

Maike Ahrens (ahrensmaike), Sebastian Voss (svoss09)