[multistage] run TPC-H benchmark on multistage engine
- Dominant language
- Java
- Stars
- 6.1k
- Forks
- 1.5k
- Avg merge
- 2d 3h
- Merged PRs (30d)
- 195
Description
currently, we are not 100% supported for TPC-H syntax. we would like to track our progress towards TPC-H here
1. run TPC-H data loading and TPC-H queries in integration test (See: #9559) - disable the ones that are not supported
- [ ] add query dbgen dataset for TPC-H
- [ ] add query with properly filled in parameters for TPC-H (we don't support parametric replacement at the moment)
2. add features and TPC-H query supported
- [ ] DateTime functions are not supported (e.g. `orderdate < date '2000-01-01' + interval '3' month`)
- [ ] some JOIN filter parsing (e.g. `(a=b AND c > 0) OR (a=b AND d < 0)` doesn't get converted to `a=b AND (c > 0 OR d < 0)`)
- [ ] logical correlate
- [ ] single-value sub query join (`WHERE EXIST (SELECT ...)`, `WHERE a > (SELECT COUNT(*) FROM ...)`)
- [ ] table view not supported
CC @ankitsultana @kishoreg @xiangfu0
Contributor guide
Research direction
Start by reviewing issue #9559 and the integration-test path for TPC-H data loading and queries. Check how the dbgen dataset and query parameters are represented, then assess each listed unsupported feature; done means the checklist is implemented and supported queries run in integration tests.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- java, sql
- Domain
- databases
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 25/100