Skip to contents

Get required input variables for a NONMEM model

get_required_input_variables.Rd

Parses a NONMEM model and determines which variables from $INPUT are required to create a new input dataset. Variables are classified as:

  • "reserved" – standard NONMEM data items with intrinsic meaning (ID, TIME, DV, AMT, EVID, etc.)

  • "dose_variable" – columns referenced on the right-hand side of a dose-timing parameter assignment (D1–D9, ALAG1–ALAG9, F1–F9, R1–R9) in $PK (e.g. D1 = DUR or D1 = DUR * 24). These must be specified per dose event, not per subject.

  • "used_covariate" – non-reserved columns explicitly referenced in the model code ($PK, $DES, $ERROR, $PRED) but not classified as a dose variable

  • "unused_covariate" – columns present in $INPUT but never referenced in model code

  • "dropped" – columns marked DROP in $INPUT

Usage

get_required_input_variables(
  model,
  data = NULL,
  include_reserved_nonmem = TRUE
)

Arguments

model

Path to a NONMEM .mod/.ctl file, NONMEM model code as a single string, or a Pharmpy NONMEM model object.

data

Optional. A data.frame or path to a CSV file whose column names are used to populate data_col. Columns are matched to $INPUT entries positionally (same order). When omitted the filename from $DATA is tried automatically; if that file cannot be found, data_col falls back to nonmem_name.

include_reserved_nonmem

Logical. If TRUE (default), reserved NONMEM variables are included in the returned data frame. Set to FALSE to return only covariate-type variables (useful when you only need to know which subject-level covariates to include).

Value

A data.frame with columns:

nonmem_name

Name used inside NONMEM model code.

data_col

Corresponding column name in the data file.

type

Classification: "reserved", "dose_variable", "used_covariate", "unused_covariate", or "dropped".

required

TRUE if the column must be present in a new input dataset.

Details

"reserved", "dose_variable", and "used_covariate" columns are all considered required for simulation. Renames in $INPUT (e.g. WT=WEIGHT, where WT is the NONMEM internal name and WEIGHT is the data-file column) are handled correctly.