prep_clean_labels: Support function to scan variable labels for applicability

Description

Adjust labels in meta_data to be valid variable names in formulas for diverse r functions, such as glm or lme4::lmer.

Usage

prep_clean_labels(label_col, meta_data = "item_level", no_dups = FALSE)

Value

a data.frame with:

if meta_data is set, a list with:
- modified meta_data[, label_col] column
if meta_data is not set, adjusted labels that then were directly given in label_col

Arguments

label_col: character label attribute to adjust or character vector to adjust, depending on meta_data argument is given or missing.
meta_data: data.frame metadata data frame: If label_col is a label attribute to adjust, this is the metadata table to process on. If missing, label_col must be a character vector with values to adjust.
no_dups: logical disallow duplicates in input or output vectors of the function, then, prep_clean_labels would call stop() on duplicated labels.

Details

Currently, labels as given by label_col arguments in the most functions are directly used in formula, so that they become natural part of the outputs, but different models expect differently strict syntax for such formulas, especially for valid variable names. prep_clean_labels removes all potentially inadmissible characters from variable names (no guarantee, that some exotic model still rejects the names, but minimizing the number of exotic characters). However, variable names are modified, may become unreadable or indistinguishable from other variable names. For the latter case, a stop call is possible, controlled by the no_dups argument.

A warning is emitted, if modifications were necessary.

Examples

Run this code

if (FALSE) {
meta_data1 <- data.frame(
  LABEL =
    c(
      "syst. Blood pressure (mmHg) 1",
      "1st heart frequency in MHz",
      "body surface (\\u33A1)"
    )
)
print(meta_data1)
print(prep_clean_labels(meta_data1$LABEL))
meta_data1 <- prep_clean_labels("LABEL", meta_data1)
print(meta_data1)
}