Skip to contents

What the data hold, for the review

Source: R/review_facts.R
tfl_data_facts.Rd

Reads the datasets once and keeps the facts tfl_review_spec() checks a definition against: every column's class, label, number of distinct values and, for a character or factor column with at most max_levels of them, the values; each dataset's rows and subjects; each analysis set's subjects and the values of its flag; and each condition's rows and subjects (or why it cannot be evaluated). The facts are small (a few KB a dataset) and hold no records, so they can be kept between sessions.

Usage

tfl_data_facts(
  data,
  populations = NULL,
  conditions = NULL,
  max_levels = 200L,
  listings = NULL,
  id = "USUBJID"
)

Arguments

data

The datasets: a named list of data frames (names as the ARD definition's datasets sheet gives them: ADSL, ADAE ...).

populations

The analysis sets: the populations sheet (population_id, dataset, where), or a whole ARD definition (tfl_ard_spec() or its sheets), whose analysis data and analyses then give the conditions too.

conditions

More conditions to count: a data frame with dataset, where and optionally population_id (the analysis set the rows are first narrowed to). The ARD definition's own are added when populations is one.

max_levels

A column with more distinct values than this keeps only their number (200, as many as a condition builder offers).

listings

A listing definition (tfl_listing_spec()): its where are counted too.

id

The subject key.

Value

A tfl_data_facts: a list of datasets (each n, n_subjects and columns, a data frame with name, class, label, n_distinct and the list column values), populations (each dataset, n_subjects, flag, flag_values and the values of the columns the definition reads among its subjects, values), conditions (a data frame: dataset, population_id, where, n_rows, n_subjects, error, and the list column values), ard (empty: what an ARD holds, tfl_ard_facts(), is added by name) and made (the time).

See also

tfl_review_spec(), tfl_ard_facts()

Examples

adsl <- data.frame(USUBJID = c("1", "2", "3"), SAFFL = c("Y", "Y", "N"),
                   SEX = c("F", "M", "F"), AGE = c(50, 61, 47))
f <- tfl_data_facts(list(ADSL = adsl),
  populations = data.frame(population_id = "SAF", dataset = "ADSL",
                           where = 'SAFFL == "Y"'))
f$populations$SAF$n_subjects
#> [1] 2
f$datasets$ADSL$columns[, c("name", "class", "n_distinct")]
#>      name     class n_distinct
#> 1 USUBJID character          3
#> 2   SAFFL character          2
#> 3     SEX character          2
#> 4     AGE   numeric          3