Selective Late Materialization in Modern Analytical Databases
Summary: Classic late-materialization loses value in vectorized columnar engines; this paper redefines LM and proposes Selective Late Materialization (SLM), choosing per-attribute, per-query materialization points to minimize cost. Implemented in DuckDB, SLM outperforms early/late materialization on JOB by 14.7%/8.9% average (up to 76.7% per query) and shows similar TPC‑DS gains. (summarized by gpt-5-mini on Feb 09 2026)
Incoming Non-self Citations Over Time
No non-self incoming citations found for this paper in this database.
Authors
- 1. Yihao Liu
- 2. Shaoxuan Tang
- 3. Yulong Hui
- 4. Hangrui Zhou
- 5. Huanchen Zhang
Incoming Citations (Sorted by Pagerank)
Showing 0 of 0 citing papers.
| Rank | Citing Paper | Year | Venue | Pagerank |
|---|
Previous
Page 1 / 1
Next
Outgoing Citations (Sorted by Pagerank)
Showing 23 of 23 cited papers.
Citations counted here include only citations to other VLDB/SIGMOD/CIDR/PODS papers in this database.
Previous
Page 1 / 1
Next
Semantically Similar Papers
| Overall Rank | Paper | Year | Venue | Pagerank |
|---|---|---|---|---|
| 1,112 | Materialized View Selection and Maintenance Using Multi-Query Optimization | 2001 | SIGMOD | 0.00013917776 |
| 7,126 | Debunking the Myth of Join Ordering: Toward Robust SQL Analytics | 2025 | SIGMOD | 4.8232367e-05 |
| 761 | Materialization Optimizations for Feature Selection Workloads | 2014 | SIGMOD | 0.00017053783 |
| 5,535 | Lightweight Cardinality Estimation in LSM-based Systems | 2018 | SIGMOD | 5.4539235e-05 |
| 8,276 | DuckPGQ: Efficient Property Graph Queries in an analytical RDBMS | 2023 | CIDR | 4.5439811e-05 |
| 8,275 | Adaptive Factorization Using Linear-Chained Hash Tables | 2025 | CIDR | 4.5439841e-05 |
| 7,427 | Selection Pushdown in Column Stores using Bit Manipulation Instructions | 2023 | SIGMOD | 4.7327406e-05 |
| 10,635 | Saving Private Hash Join | 2025 | VLDB | 4.1945683e-05 |
| 9,504 | Supporting Scalable Analytics with Latency Constraints | 2015 | VLDB | 4.3341665e-05 |
| 10,372 | Data Chunk Compaction in Vectorized Execution | 2025 | SIGMOD | 4.1945683e-05 |