Apply dataset-specific value corrections before cleaning

Rewrites cell values in named columns using a small pattern/replacement table. This is the escape hatch for dataset quirks the reference lists cannot express: a "Total " prefix on every analyte, a known-bad CAS, a unit typo. Runs before the cleaning pipeline so every downstream step sees the corrected value.

Usage

apply_value_corrections(df, value_corrections = NULL)

Arguments

df

Data frame after frontmatter extraction and janitor::clean_names().

value_corrections

Data frame with columns column, pattern, replacement, and optional match_mode (one of "regex" (default), "literal_exact", or "literal_word"). Rows are applied in order.

Value

List with cleaned_data and audit_trail (step value_correction).