Usage¶

Open and Append Records¶

from sdsavior import SDSavior

rb = SDSavior("data.ring", "data.meta", 8 * 1024 * 1024)
rb.open()
try:
    rb.append({"event": "boot"})
    rb.append({"event": "sample", "value": 123})
finally:
    rb.close()

Context-manager usage is recommended:

with SDSavior("data.ring", "data.meta", 8 * 1024 * 1024) as rb:
    rb.append({"ok": True})

Read Records¶

for seq, ts_ns, obj in rb.iter_records():
    print(seq, ts_ns, obj)

Filter by sequence:

for seq, ts_ns, obj in rb.iter_records(from_seq=200):
    ...

Export JSONL¶

with SDSavior("data.ring", "data.meta", 8 * 1024 * 1024) as rb:
    rb.export_jsonl("out/export.jsonl")

Durability Controls¶

fsync_meta=True (default): metadata is fsync'd after writes.
fsync_data=False (default): data pages are not fsync'd on every append.
recover_scan_limit_bytes=None (default): scan up to capacity during recovery.

Use fsync_data=True when stronger durability is required and throughput tradeoffs are acceptable.