import neo4j
driver = neo4j.driver("bolt:https://localhost")
session = driver.session()
session.run("CREATE (a:Person {name:'Bob'})")
for name, in session.run("MATCH (a:Person) RETURN a.name AS name"):
print(name)
session.close()python -m neo4j "CREATE (a:Person {name:'Alice'}) RETURN a, labels(a), a.name"$ python -m neo4j.bench -p 8 "MATCH (a:Person {name:'Alice'}) RETURN a"
Neo4j Benchmarking Tool for Python
Copyright (c) 2002-2015 "Neo Technology,"
Network Engine for Objects in Lund AB [http:https://neotechnology.com]
Report bugs to [email protected]
Python 3.4.1 (default, Nov 27 2014, 09:15:30)
[GCC 4.8.2]
This machine has 8 processors (linux)
Latency measurements:
INIT <---SEND---> <---RECV---> DONE
overall <--------------------------------------------->
network <--------------------------->
wait <--->
INIT - start of overall request
SEND - period of time over which data is sent
RECV - period of time over which data is received
DONE - end of overall request
Using graph database at bolt:https://localhost
Warming up... done
------------------------------------------------------
MATCH (a:Person {name:'Alice'}) RETURN a
× 10,000 runs × 8 clients = 2,601.3 tx/s
(80,000 requests in 30.8s)
------------------------------------------------------
percentile | overall | network | wait
------------|-------------|-------------|-------------
0.0% | 1,370.2µs | 199.0µs | 9.9µs
10.0% | 2,673.3µs | 244.6µs | 18.3µs
20.0% | 2,750.5µs | 270.2µs | 19.9µs
30.0% | 2,790.2µs | 319.1µs | 21.9µs
40.0% | 2,839.1µs | 331.8µs | 180.4µs
50.0% | 2,921.7µs | 343.7µs | 197.1µs
60.0% | 3,039.7µs | 355.8µs | 217.1µs
70.0% | 3,106.4µs | 368.1µs | 283.9µs
80.0% | 3,176.3µs | 382.8µs | 303.4µs
90.0% | 3,334.3µs | 404.0µs | 329.8µs
95.0% | 3,633.0µs | 541.3µs | 396.6µs
98.0% | 4,888.2µs | 1,565.6µs | 954.5µs
99.0% | 5,767.4µs | 2,585.5µs | 2,049.7µs
99.5% | 6,810.7µs | 3,172.1µs | 2,846.9µs
99.9% | 10,824.5µs | 5,846.7µs | 4,604.3µs
100.0% | 29,740.3µs | 25,134.4µs | 25,065.5µs./runprofile.sh "RETURN 1" | less