You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Commit 8d75270
Browse filesBrowse the repository at this point in the historyBrowse files
perf(snowflake): cut per-model statement overhead, and write a Rules-failed compile to target/compiled-failed
- Drop and replace only the table kinds a build creates.
- Snowflake: per-run `SHOW` schema listings and retention; session-only statements such as `QUERY_TAG` keep them.
- Fingerprint rows are appended only when the identity changes. Skipping is off with model migration history: rename matching reads fingerprint times.
- Scenario cleanup lists targets by name prefix and reuses confirmed results.
- Kept: merge/microbatch column reads (warehouse types), the DuckDB staging drop after a swap, the freshness `LAST_ALTERED` read, the `sqb test` path.
- A compile whose Rules fail writes its SQL to `target/compiled-failed/` (with the not-buildable header) and never touches `target/compiled/`. The next passing compile deletes it, and the report names each failing file. `target/compiled-models.txt` is removed.
- The compile after a Rules-failed compile runs in full, since that compile writes no SQL tests, functions or audits.
- Goldens:
- `failure/rule-custom-finding.json`: model SQL moves from `compiled` to `compiled_failed`.
- `failure/rule-unused-cte.json`: same, with the header.
- `failure/rule-unused-cte-after-variable-and-macro.json`: same, with the header.
0 commit comments