GX Core (https://github.com/fivetran/great_expectations 11,856 ) (Apache-2.0, formerly Great Expectations 385,595 ) expresses each check as a declarative expectation, such as ExpectColumnValuesToBeUnique(column="order_id"). An expectation suite groups them; a batch definition names the data (a pandas 16,086 or Spark 129 DataFrame, or a SQL table); a validation definition pairs the two; and a checkpoint runs validations and triggers actions such as rebuilding the Data Docs HTML site. Results carry observed values and samples of unexpected rows, and mostly=0.999 turns any expectation into a tolerance.
GX Core 1.x requires Python below 3.14, so here it runs in a Python 3.12 environment (uv venv -p 3.12). Its ownership changed in 2026: on 6 May the company announced that FICO had acquired GX Cloud (withdrawn from public sale on 1 June) and that Fivetran 16,237 became steward of GX Core, whose repository now sits under fivetran; releases continue (1.23.2, 25 September 2026).