Thanks for the clear repro details — this is a known issue with Athena when identifiers (catalog/schema/table) include hyphens and the DROP statements aren’t quoted. What we know from the knowledge sources: - Similar Athena failures occur when Elementary (or its cleanup step) issues unquoted DROP TABLE statements; quoting with backticks or adapter-aware quoting resolves the parse error (mismatched input ‘-’) in Athena ([Slack: schema with dash](https://slack.com/archives/C02CTC89LAX/p1709655251458079)). - The team addressed related Athena cleanup bugs where temporary test tables weren’t fully dropped; the fix introduced adapter-specific cleanup via a fully_drop_relation macro and Athena-specific clean-up, acknowledging Athena’s different behavior for temp tables and file cleanup ([GH #1514](https://github.com/elementary-data/elementary/issues/1514)). - For quoting mismatches in Snowflake/Databricks, Elementary provides a CLI flag --dbt-quoting to override quoting for its internal package, but that applies when running edr commands, not dbt test hooks ([Slack: dbt-quoting flag](https://slack.com/archives/C02CTC89LAX/p1700141940747659), [Slack: on_run_end quoting](https://slack.com/archives/C02CTC89LAX/p1729840602042949)). - Users have worked around Databricks/Athena quoting issues by overriding package macros to enforce proper quoting/backticks for catalogs with special characters ([Slack: Databricks backticks](https://slack.com/archives/C02CTC89LAX/p1696522119299639)). Recommended path forward: 1) Confirm you’re on the latest Elementary dbt package version that includes the Athena cleanup fixes (the fix in GH #1514 was merged). If you’re already on 0.19.4 and still seeing unquoted DROP, it suggests a remaining edge case in the DROP path for Athena ([GH #1514](https://github.com/elementary-data/elementary/issues/1514)). 2) Short-term workaround: override the relevant Elementary macro in your dbt project to enforce adapter-aware quoting for DROP. The pattern used in prior workarounds is to build a Relation via the adapter and render it quoted, or to ensure the catalog/schema are quoted/backticked per adapter rules (similar to the Databricks override example) ([Slack: Databricks backticks](https://slack.com/archives/C02CTC89LAX/p1696522119299639), [Slack: schema with dash](https://slack.com/archives/C02CTC89LAX/p1709655251458079)). 3) If feasible, run a cleanup macro using an adapter-specific fully_drop_relation that calls both adapter.drop_relation and Athena’s clean_up_table to remove catalog metadata and S3 files for temp tables (this pattern was introduced for Athena) ([GH #1514](https://github.com/elementary-data/elementary/issues/1514)). 4) If you’re not invoking edr, the --dbt-quoting flag won’t help for dbt test hooks. The fix needs to be in the package macro path that generates DROP statements during test cleanup ([Slack: on_run_end quoting](https://slack.com/archives/C02CTC89LAX/p1729840602042949)). If you can share a minimal debug log snippet pointing to the exact macro path generating the DROP (e.g., clean_elementary_test_tables or similar), we can confirm where to place the override. If this persists on 0.19.4 with dbt-core 1.10.9, it’s likely a remaining bug — please open an issue and we’ll prioritize an adapter-specific quoting fix for Athena’s DROP path (and happy to review a PR if you have one). ([Slack: schema with dash](https://slack.com/archives/C02CTC89LAX/p1709655251458079), [GH #1514](https://github.com/elementary-data/elementary/issues/1514)).