Time Travel

• Cypher
• Python SDK

// Show commit history
CALL db.history()

// Read a previous commit
checkout c0a6ce3c0315be7a

// Query graph at specific commit
MATCH (n:Person)-[e:KNOWS]->(m:Person)
WHERE n.name = 'Jane'
AND m.name = 'John'
RETURN n, e, m

# Show commit history
client.query("CALL db.history()")

# Read a previous commit
client.checkout(change="main", commit="c0a6ce3c0315be7a")

# Query graph at specific commit
client.query("""
MATCH (n:Person)-[e:KNOWS]->(m:Person)
WHERE n.name = 'Jane'
AND m.name = 'John'
RETURN n, e, m
""")

​

Lazy Commit Loading

TuringDB lazily loads commits from disk. When a graph is loaded, only the HEAD commit (the latest commit on the current branch) is fully loaded into memory. All other commits are registered but their data remains on disk until explicitly requested.

This means that to query a past commit, its data must first be loaded into memory using the LOAD COMMIT command:

LOAD COMMIT 'c0a6ce3c0315be7a'

​

Example: Querying a Past Commit via the REST API

// Step 1: Load the commit into memory
LOAD COMMIT 'be9643a1c03f8e2b'

// Step 2: Now you can query at that commit
MATCH (n:Person)-[e:KNOWS]->(m:Person)
RETURN n.name, m.name

LOAD COMMIT is idempotent — calling it on an already-loaded commit is a no-op and returns immediately.