Mini SQL query optimizer

Input: a SQL query and a JSON catalog with schema statistics. Output: the cheapest evaluation plan, printed as an operation tree with the chosen algorithm and cost (in block transfers) per operation.

Supported SQL subset: SELECT attributes (or *) from up to 4 tables, WHERE as a conjunction of up to 6 conditions, ORDER BY one attribute.

Python, no dependencies (pytest only for tests).

Usage

python main.py <catalog.json> "<SQL query>"
python main.py <catalog.json> <query.sql>

Example:

python main.py tests/fixtures/primer_ulaza.json "SELECT ime FROM Student WHERE indeks = '2020/1234'"

Tests

pytest tests/ -v

Layout

• src/catalog/ - schema model and JSON catalog loader
• src/sqlparser/ - tokenizer, AST and recursive descent parser
• src/semantic/ - name resolution, selection vs join classification, type checks
• src/cost/ - cost formulas: selectivity, selection algorithms A1-A9, external sort, joins
• src/optimizer/ - dynamic programming join enumeration and planner
• src/plan/ - plan tree and printer
• main.py - CLI entry point

Notes on scope, conventions and verification against course materials (lecture slide example, exam solutions, official examples folder): NOTES.md.