Back to changelog

Spreadsheet Parsing Bug Fixes

Fixes to how .xlsx and .xlsm workbooks are parsed. They apply to the Parse API and to spreadsheet indexing. Rerun a parse job, or reindex a file, to pick them up.

  • Dropdown choices are kept: every choice of a dropdown now appears in the output, not only the selected value. Each dropdown is one line in the chunk that holds its cell, for example - H26 "Packaging included:": dropdown: None | Returnable | Expendable | Mix (selected: Returnable).
  • All list sources: choices are read from typed lists, cell ranges, other sheets, named ranges and Excel Tables. A dropdown that depends on another cell lists the choices for that cell's current value.
  • Other input rules: number, date, time and text-length rules come out as text, such as whole number between 1 and 100, along with any input message.
  • Long lists say when they are cut: a list longer than 1,000 choices shows the first 1,000 and says the list has more.
  • Merged cells beside rows: a value merged down beside many rows, such as one objective next to every step of a process, no longer turns the first rows into the header. Those rows keep all their cells, and the table is no longer split part way down.