Database Change Detection Without Triggers

Try it yourself — zero dependencies

pip install atomik-core
python -m atomik_core.examples.sqlite_change_tracker

Creates a 1000-row SQLite database, modifies 50 rows, and detects every change using XOR fingerprinting. No external dependencies — SQLite is built into Python.

How it works

from atomik_core import Fingerprint, AtomikTable
import sqlite3

# 1. Connect to your database
db = sqlite3.connect("myapp.db")
tracker = AtomikTable(num_contexts=65536)

# 2. Fingerprint each row (initial snapshot)
for row in db.execute("SELECT id, name, email, balance FROM users"):
    fp = xor_reduce(serialize(row))  # 8-byte fingerprint
    tracker.load(addr=row[0], initial_state=fp)

# 3. On next sync cycle — detect changes in O(1) per row
changed_ids = []
for row in db.execute("SELECT id, name, email, balance FROM users"):
    fp = xor_reduce(serialize(row))
    if tracker.read(addr=row[0]) != fp:
        changed_ids.append(row[0])
        tracker.load(addr=row[0], initial_state=fp)

print(f"{len(changed_ids)} rows changed")
# Only sync/replicate the changed rows

Comparison