Rbbt uses TSV (tab-separated value) files as a first-class data format.

The tooling focuses on:

  • large files (millions of rows)
  • streaming (don’t load everything into memory)
  • semantics (columns and “value types” are meaningful)
  • indexing / persistence (fast random access)

TSV “types” (value shapes)

Historically, Rbbt classifies TSVs by the shape of the value column(s):

  • single: one value per key
  • list: list of values per key
  • double: list-of-lists per key (multiple columns, each a list)
  • flat: list of values representing the same aspect (e.g. transcripts)

When to use it

  • Interchange format between tasks
  • Cached artifacts (fast to inspect, easy to diff)
  • Lightweight data tables for workflows

Deeper docs

This site uses Just the Docs, a documentation theme for Jekyll.