Only registered operations with typed parameters reach the engine. Rules that delete data (drop rows, duplicates) are always unchecked by default.
trim_whitespace
Removes leading and trailing spaces from text. Never touches numbers or formulas.
normalize_spaces
Collapses repeated inner spaces into one. Preserves tabs and line breaks.
remove_control_characters
Strips specific invisible control characters (e.g. BOM, zero-width). Nothing beyond the listed ones.
normalize_case
Standardizes case: lower, UPPER or Title. Optional; does not fix spelling.
normalize_email
Trims edges and lowercases the domain. Does not repair an invalid email.
normalize_date
Converts dates to one format (DD/MM/YYYY or YYYY-MM-DD). Ambiguous ones are left as is.
normalize_phone
Validates and formats BR phone numbers (national or E.164). Never invents area codes or digits.
normalize_cpf
Validates CPF format and check digits before masking. Invalid CPFs stay intact.
normalize_cnpj
Validates and standardizes CNPJ in supported formats. Uncovered formats are left as is.
normalize_currency
Parses currency with explicit locale and Decimal. No FX, summing or precision loss.
normalize_boolean
Maps value lists to true/false (e.g. Yes/No). Unknown values stay put.
map_values
Standardizes many spellings in a column into one canonical value (e.g. SP/São Paulo → SP).
replace_null_values
Fills only recognized empties/nulls with a value you choose. No inference.
rename_column
Renames only the chosen header, with no name collisions.
convert_type
Strictly converts type (text/integer/decimal), protecting zeros and precision.
remove_empty_rows
Removes only fully empty rows. A formula does not count as empty.
remove_duplicates
Removes exact duplicates on key columns, keeping the first or the last.
drop_rows_missing_required
Removes rows where a required field is empty. Optional (opt-in).