@dongfengweixiao @bazylhorsey I agree with both of you that using an Id like a snowflake or another timestamp base Id could be beneficial for better ordering, improving indexing, reducing the database size, and potentially reducing id generation time if it is less than 128 bits and maybe the created_at column can be removed. Currently, I am doing some research and testing so in a next commit this will be implemented.

References:
https://medium.com/geekculture/the-wild-world-of-unique-identifiers-uuid-ulid-etc-17cfb2a38fce
https://medium.com/double-pointer/system-design-interview-scalable-unique-id-generator-twitter-snowflake-or-a-similar-service-18af22d74343
https://itnext.io/choosing-the-right-data-type-means-of-generating-unique-primary-keys-d7aac92968c6
https://blog.devgenius.io/analyzing-new-unique-identifier-formats-uuidv6-uuidv7-and-uuidv8-d6cc5cd7391a
https://www.ietf.org/archive/id/draft-peabody-dispatch-new-uuid-format-04.html

Implementations
https://github.com/agnirudrasil/avault/blob/c6ca7037a5a99d97fcf43eb6b9ef935439171c02/backend/app/api/utils/snowflake.py
https://github.com/SawdustSoftware/simpleflake/blob/master/simpleflake/simpleflake.py
https://github.com/sony/sonyflake
https://github.com/ergeon/python-flax-id
https://github.com/AmreeshTyagi/goldflake
https://github.com/oittaa/uuid6-python
https://github.com/uuid6/uuid6-ietf-draft

May be a good idea to use snowflake as the primary key? #12

Description

Metadata

Metadata

Assignees

Labels

Projects

Milestone

Relationships

Development

Issue actions