Overview
Intelligent Index Builder is a team project developed during November–December 2024. It recommends and constructs indexes for relational databases and has obtained software copyright registration.
Workload analysis
JSQLParser parses SQL queries. EXPLAIN plans and information_schema statistics provide cardinality and selectivity information for estimating candidate index benefits.
Recommendation
Greedy pruning and cost estimation select beneficial index combinations while mitigating exponential growth in the candidate space. Recommendations can be updated dynamically.
Construction and feedback
Indexes are created with LOCK=NONE. After construction, EXPLAIN compares execution costs before and after; indexes without a significant benefit are removed.
My contribution
Contributed SQL parsing, workload analysis, benefit estimation, greedy pruning and post-construction feedback as part of the team.