Introduction
The Standard runtime is Groundzero's default, zero-infrastructure way to turn a connection into a dataset. You write one read-only query in the connection's own dialect, optionally add a few lines of pandas "Prep" code, and Groundzero returns pages of rows. You size no cluster, start no session and import no helper library.
What \"Standard\" means in Groundzero
The word covers three related things, and this guide keeps them apart.
- Standard datasets (the main subject of this guide): the Standard choice in the dataset designer. The query runs in Groundzero's managed query engine, not on a runtime you start.
- The Python (pandas) job runtime: a runtime for jobs and code datasets, for plain Python and pandas workloads.
- The base runtime: the Python 3.12 foundation that the Spark and DuckDB runtimes, and every custom runtime, are built on.
Spark and DuckDB each ship a Groundzero helper library for use in code. Standard has none: its programming surface is SQL plus the Prep contract described in the API reference.