break large row to many pack to lower scan latency
Open
stat/will-see-later
- Dominant language
- C++
- Stars
- 1.9k
- Forks
- 432
- PR merge metrics
- No merged PRs in 30d
Description
low-level-scan中的filter以row为单位进行，因此每次scan rpc的最小粒度是一行，对于超大行来说延迟非常大。
对于用户没有指定filter的情况，这种做法没有必要，一行可以被分解到多个rpc。
Contributor guide
No contributing guide indexed for this repository
Research direction
Start by tracing the low-level-scan path and its row-based filter handling, then identify how scan RPC boundaries are formed when no filter is supplied. Verify the behavior with a large row and confirm that it can be divided across multiple RPCs while reducing scan latency.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- cpp
- Domain
- databases, performance
- Issue type
- Feature
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 35/100