RocksDB history and transactions

Choose history retention when opening a store. The flags control which physical column families Maryk maintains; see the storage layout.

History retention

• keepAllVersions defaults to true, retaining historic table, index and unique values for time-travel and change-history reads.
• keepUpdateHistoryIndex defaults to false. Enable it to maintain the model-wide index keyed by change version and record key for historical paging. ScanUpdates may select records through this index; its synchronization events remain chronological.

For example, keep both history structures for a model named Account:

import maryk.datastore.rocksdb.RocksDBDataStore

val store = RocksDBDataStore.open(
    relativePath = "accounts.db",
    dataModelsById = mapOf(1u to Account),
    keepAllVersions = true,
    keepUpdateHistoryIndex = true,
)

Close the store after use. See the module overview for a complete opening and cleanup example.

Revision paging

For finite, newest-first revision pages, use getUpdateHistory(key, ...) or scanUpdateHistory(...) with both history options enabled and a ready history index. Continue with UpdatesResponse.nextCursor; the limit counts complete record-version entries, independently of property-history maxVersions. Both queries scan the existing model-wide history index. See history paging for filters, projections, and hard-delete retention limits.

Hard deletion

A hard delete permanently removes a record’s current and historic values, including when keepAllVersions is enabled. An earlier toVersion query cannot recover erased values. Use soft deletion when historical recovery is required.

See store behavior for undelete, unique ownership, replication ordering and scans that include deleted records.

Snapshots and reopen

Committed atomic requests and transactions persist their version watermark with their data and receipt, so snapshots remain current after reopening. See opening existing stores for the one-time watermark repair on older stores.

Serializable transaction scopes

RocksDB replays through a native optimistic transaction. Point reads and iterators include its pending writes; a revision key participates in conflict detection and changes in each normal write commit. Preview closes without commit, and durable receipts survive reopen.

The shared contract describes read-your-own-writes, savepoints, conservative revision conflicts, deadlines and commit resolution. Query and mutation APIs shipped in 0.12.1 retain their existing contracts outside an explicit transaction.