Assigns functional roles (e.g. predictor, outcome, identifier, time variable,
strata, weights) to variables across a single data frame or an entire R database list.
Produces standardized schemas compatible with validate_column_mapping,
recipes, and modeling workflows.
Usage
variable_roles(
data,
id = NULL,
outcome = NULL,
time = NULL,
predictors = NULL,
strata = NULL,
weight = NULL,
ignore = NULL,
roles = NULL
)
set_roles(data, ...)
get_roles(data)
get_predictors(roles)
get_outcomes(roles)
get_id_cols(roles)
get_time_cols(roles)Arguments
- data
A data frame or a named list of data frames (an R database).
- id
Character vector of column names acting as subject/patient/cluster identifiers.
- outcome
Character vector of column names acting as response/outcomes (
Y_var).- time
Character vector of column names acting as longitudinal visit/time variables (
Time_var).- predictors
Character vector of column names acting as features/predictors (
X_var).- strata
Character vector of column names used for stratification or subgroup analysis.
- weight
Character vector of column names used for case weights or survey weights.
- ignore
Character vector of column names to ignore/exclude from modeling.
- roles
Role table (e.g. from
variable_roles) or data object.- ...
Additional arguments passed to
variable_roles.
Value
A tibble with columns columns, role, X_var, Y_var,
ID_var, Time_var, and dataset_name (when multi-table).
Examples
df <- data.frame(id = 1:5, time = 1:5, sbp = c(120, 130, 115, 140, 125), death = c(0, 0, 1, 0, 1))
roles <- variable_roles(df, id = "id", time = "time", outcome = "death", predictors = "sbp")