Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Bloom filter use pk_prefix instead of distribution_key #6825

Closed
1 of 2 tasks
wcy-fdu opened this issue Dec 9, 2022 · 1 comment · Fixed by #6871
Closed
1 of 2 tasks

Bloom filter use pk_prefix instead of distribution_key #6825

wcy-fdu opened this issue Dec 9, 2022 · 1 comment · Fixed by #6871
Assignees

Comments

@wcy-fdu
Copy link
Contributor

wcy-fdu commented Dec 9, 2022

Perviously, we use distribution_key as our bloom filter key, and distribution_key is always the prefix of pk. After watermark design, we can not ensure the distribution key be the prefix of pk(#6288), so #6575 has changed the prefix_hint to dist_key_hint for bloom_filter.
However, using distribution key as bloom filter key will make things more complex, for example: we need to handle many corner cases, such as shuffled distribution key and discontinuous distribution key. And we need to intercept distribution key from pk, and do many judgement in StateTable/StorageTable/FilterKeyExtract.

After some discussion, we can decouple bloom filter key and distribution key, just use pk prefix as the bloom filter key, this makes things easier and bloom filter will be used in more places.

  • add prefix_len in catalog
  • change bloom filter key from distribution key to pk prefix
@github-actions github-actions bot added this to the release-0.1.15 milestone Dec 9, 2022
@wcy-fdu wcy-fdu self-assigned this Dec 9, 2022
@wcy-fdu
Copy link
Contributor Author

wcy-fdu commented Dec 9, 2022

cc @hzxa21 @Li0k

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

Successfully merging a pull request may close this issue.

2 participants