Closed
Milestone
Jan 1, 2023–Mar 31, 2023
2023, Q1
Goals
- Implement statistics subsystem in the Tarantool core (can be carried over to Q2)
- Implement distributed aggregates (no custom aggregate functions)
- Implement a new Picodata engine
- "Not" support
Results
Most of the goals were unfinished and still are in progress (carried over to Q2). I believe that the engine refactoring (storage runtime now uses tarantool spaces for temporary tables) is the reason of the broken deadlines. We got lots of bugs after this refactoring.
The good news are: we started PG protocol (Q3) and refactored the tree traversal (Q4).
Unstarted Issues (open and unassigned)
0
Ongoing Issues (open and assigned)
0
Completed Issues (closed)
42
- possible wrong README for stress-test
- bug: incorrect motion policy for operators `!=`, `not in`
- tarantool does not remove subfolders
- flaky insert error `data did not match any variant of untagged enum EncodedValue`
- Research: how COIO works under hood
- bug: generated column names may match column names generated by tarantool
- Describe the logic of statistics application algorithm during IR plan cost estimation.
- flaky integration test when cluster is replicated
- Add new engine trait for statistics
- Implement algorithm of string distance calculation from PostgreSQL
- Support statistics methods for the Value structure
- Implement histogram structures
- Implement COUNT for a column
- bug: virtual table has tuple with wrong type
- Slow grammar parsing performance on select_cast_plan_nested_where2 test
- Fix multiple key in segment distribution for join
- incorrect motion policy for inner join
- sbroad fails to execute query with nested inner joins
- sbroad fails to plan query with nested subqueries
- Do not try to prepare DML queries
- Fix Motion(Segment) with Buckets::All
- Failed to create PK for the virtual table
- Panic with already started transaction
- Query with params error
- Borrowing runtime error
- Support GROUB BY for arbitrary expression
- Temporary spaces for virtual tables on the storages
- Support arbitrary expressions with arithmetics in projections
- Support arbitrary expressions with arithmetics in selections and joins
- Flaky integration on CI: target_queries.test_type_23
- Research: histogram statistics transformation in complex queries
- Research: PostgreSQL protocol support
- Divide integration tests to two clusters
- Flaky test on sbroad.execute("select")
- Mock histogram on router
- Fix anonymous column names
- New API for the sbroad errors
- Research Sbroad migration to Picodata (from Cartridge)
- Improve locality for INSERT SELECT
- Perfomance: replace traversal crate with custom tree walkers
- Implement GROUP BY for columns
- Добавление алгебраических операций (сложение, умножение и т.д.)
Loading
Loading
Loading