uqid.py

uqid

pip install uqid
from uqid import uqid, digits

print(uqid()) # e7ljCIs4grYZ9tOv6bpX3TQIA7FyA23umzGjDcIjHff2XwF2ziAw56B7S3EdOuCA
print(uqid(4, digits)) # 4982

• Returns unique id from random chars.
• Default is 64 chars 0-9a-zA-Z aka base62.

dtid

pip install uqid
from uqid import dtid
print(dtid())

YYYYMMDDhhmmssmicrosrnd4
20161231235959123456uN1q

• 24 bytes by default - as str(ObjectId()).
• But human-friendly - unlike ObjectId.
• Microsecond-precise UTC timestamp - unlike second-precise ObjectId().
• May extract datetime_from_dtid().
• May be used as a log timestamp as is.
• Alphabetically ordered.
• Safe chars 0-9a-zA-Z - unlike binary ObjectId() leading to "ValueError: Invalid UTF-8 sequence length" if added to JSON without str().
• base62^4 is over 14M uniques per microsecond.
• dtid(24) is 2x faster than uqid(24) but slower than str(ObjectId())

tests

python2 uqid.py
python3 uqid.py

SAMPLES:
str(uuid4()): 3bb25137-2422-4d06-bd12-5a9de2ea2bf6
len(^): 36
len(v): 24
str(ObjectId()): 5acdcd1f9c591516ae800615
uqid(24): UFRFQVCtYu6rtQq8LD33XjI3
dtid(24): 20180411085351268116tHse
Iterations: 1000000

python2 SECONDS:
str(uuid4()) seconds: 11.825964
str(ObjectId()) seconds: 2.424702
uqid(24) seconds: 9.701306
dtid(24) seconds: 5.357983

python3 SECONDS:
str(uuid4()) seconds: 5.826098
str(ObjectId()) seconds: 3.192565
uqid(24) seconds: 17.769653
dtid(24) seconds: 9.588677

DUPLICATES:
str(uuid4()) duplicates: 0
str(ObjectId()) duplicates: 0
uqid(24) duplicates: 0
dtid(24) duplicates: 0

uqid.py version 0.3.0
Copyright (C) 2015-2018 by Denis Ryzhkov denisr@denisr.com
MIT License, see http://opensource.org/licenses/MIT