Every ingestion cycle either creates the Silver table or runs a real MERGE against it:
matched earthquake IDs get updated in place, new ones get inserted — the same semantics as
MERGE INTO in Databricks SQL, just via the open-source delta-rs engine instead
of a Spark cluster. Every commit is an immutable version you can query with VERSION AS OF.
| Version | Operation | Timestamp |
|---|---|---|
| v70 | VACUUM END | 2026-09-10 00:30:14 UTC |
| v69 | VACUUM START | 2026-09-10 00:30:14 UTC |
| v68 | DELETE | 2026-09-10 00:30:14 UTC |
| v67 | MERGE | 2026-09-10 00:30:14 UTC |
| v66 | VACUUM END | 2026-09-09 08:56:18 UTC |
| v65 | VACUUM START | 2026-09-09 08:56:18 UTC |
| v64 | DELETE | 2026-09-09 08:56:18 UTC |
| v63 | MERGE | 2026-09-09 08:56:18 UTC |
No magnitude revisions detected between the last two versions.
| Date & Time (UTC) | Place | Region | Magnitude | Significance | Tsunami | Depth (km) |
|---|