The ways a TOC's output id can be made, scored

Each rule of toc_id_rules (an id column as it is; normalised to the set's id_pattern; a type column and a number column; the number in the first title, "Table 14.1.1") is tried on the rows: the one that gives unique, non-empty ids for the most rows is proposed (on a tie, an id column as it is, unless its ids carry no type and the normalised ones do).

Usage

toc_id_candidates(rows, rule_set, type_from = NULL)

Arguments

rows

What toc_rows() returned.

rule_set

A rule set.

type_from

Where the type is read ("column", "id", "title"); by default the set's, else the source that reads the most rows.

Value

A data frame, the proposed first: rule, ok (rows with a unique id), filled (rows with an id), preview (the first ids), labels (the labels they print).