Divides a row into fields at each unquoted pipe

All 8 scenarios hold

These rules cover how one line divides into fields. They apply before the parser reads a field as a value.

Each Row below is parsed as the only data row under a one-column header. Fields shows the captured field list, printed back out. Fields is text rather than a list column because a list cannot hold a null element: writing [a, , c] is itself a parse error. A null field is what several of these rules are about.

A leading pipe, a trailing pipe, or two pipes together leave a field with no text. The parser captures that field as null.

Spacing around a pipe changes nothing. One row carries both spellings of the same three fields, in place of two rows with one expectation between them.