Great Expectations 1.18.2 - Spark Connect Fix
Great Expectations 1.18.2 was published on June 26, 2026, with a targeted Spark Connect compatibility fix for Spark distinct value metrics. The release also clears BigQuery and Python 3.13 maintenance friction, fixes scheduled CI noise, and updates the docs build stack. It is a small patch release, but it touches the kind of edges that show up in real data validation work.
The full release notes and downloads are on the GitHub release page.
Spark Connect compatibility for distinct values
The main user facing fix is in Spark distinct value metrics. The notes say the project removed .rdd usage for Spark Connect compatibility in that path, covered by PR #11922.
That detail matters if Great Expectations is running checks near newer Spark deployments where Spark Connect is part of the access pattern. Code that reaches for .rdd can be a bad fit there, because Spark Connect does not expose the same direct RDD access model. This release keeps the change narrow: it targets distinct value metric handling, not a broad Spark rewrite.
For teams using Spark data sources, this is the change to test first. Run expectations that depend on distinct values against the same Spark Connect path used in production. If they failed due to .rdd access before, 1.18.2 is the patch to try.
BigQuery and Python 3.13 cleanup
The release also fixes a BigQuery collection error tied to a NumPy generic unit DeprecationWarning under Python 3.13, in PR #11924. This is maintenance work, but it is not just cosmetic for users running a newer Python runtime.
Deprecation warnings become a problem when they break tests or turn into errors in strict environments. Here the scope is BigQuery and Python 3.13. If an upgrade plan includes Python 3.13, this patch removes one known rough edge from the Great Expectations side.
CI and test suite maintenance
Great Expectations 1.18.2 includes a pytest deprecation fix around parametrize and collection handling, covered by PR #11921. The notes say this was breaking scheduled CI.
This is internal project maintenance, so most users should not expect runtime behavior to change from it. It still matters for operators who track dependency freshness and test reliability. A project with noisy scheduled CI is harder to trust when a real failure lands.
Documentation and docs build updates
The docs side gets a typo fix in the Run a Validation Definition guide via PR #11920, plus dependency bumps in /docs/docusaurus: http-proxy-middleware from 2.0.9 to 2.0.10, webpack-dev-server from 5.2.3 to 5.2.5, and @babel/core from 7.28.6 to 7.29.6.
These are not API changes. The useful point is that the published docs and docs build process stay current while the release keeps the application surface stable. The guide typo fix is small, but validation docs need precise language. One bad word in a data validation guide can waste more time than it deserves.
Upgrade notes
The notes do not call out breaking changes, deprecations, or migration steps. Treat 1.18.2 as a focused patch update. Spark and BigQuery users should run the relevant validation suites before rolling it out, especially if they depend on Spark Connect or Python 3.13.
Where to get it
- Release page: GitHub release page
- Repository: great-expectations/great_expectations
- Tag:
1.18.2