Skip to content
SiftSift

Documents

[organize]
strategy = "documents"
template = "{author}/{year}"

Placeholders

Section titled “Placeholders”
{author}
{title}
{year}
{month}
{day}

Same placeholder names regardless of whether the file is a PDF or an Office document — Sift normalizes both into the same shape before the template ever sees it.

Two formats, one strategy

Section titled “Two formats, one strategy”
  • PDF — reads the document’s /Info dictionary (Author, Title, CreationDate) directly.
  • Office (docx/xlsx/pptx) — reads docProps/core.xml (the same part in any modern Office file) for dc:creator, dc:title, and dcterms:created.

Neither requires an external tool (no LibreOffice, no Office install).

Missing metadata is a skip, never a guess

Section titled “Missing metadata is a skip, never a guess”

A PDF with no /Info, or an Office file missing docProps/core.xml, is valid but has no metadata — treated the same as any file missing the specific field its template needs: skipped with a clear reason, never given an invented value. A genuinely corrupted/unreadable file is a real error, not a “missing metadata” skip.

Recursion

Section titled “Recursion”

documents does not support --recursive — see Recursive Organization for why (shared by all four metadata strategies).

Section titled “Related”