Skip to content

timeseries-table-format

Build local time-series tables with immutable Parquet segments. The Python API tracks coverage, rejects overlapping appends, updates selected row values, and queries your tables with DataFusion SQL.

Project status

This project is an early MVP. APIs and on-disk layouts may change before v1.0.

Start here

Follow these pages in order:

  1. Check whether the project fits your workload.
  2. Install and verify the Python package.
  3. Create, append, and query your first table.

The basic workflow

  1. Create a table root with an ordered index, index granularity, and optional entity columns.
  2. Append Parquet files as new segments. The table rejects duplicate or already committed identity and index interval pairs.
  3. Register the table in a Session and query it with SQL. Results are returned as pyarrow.Table objects.

After the first-table tutorial, continue with the Common tasks section in the navigation.