Materialized View Selection in Data Warehousing

Data warehousing underpins decision support across industries by consolidating vast volumes of historical and transactional data into integrated repositories. Materialized views—stored query results computed in advance—are central to accelerating analytical workloads, yet they consume considerable storage and incur ongoing maintenance costs. The challenge of selecting which views to materialise hinges on balancing query performance against storage overhead and update latency. Optimisation frameworks typically employ cost models that quantify query execution, storage consumption and refresh operations to guide selection algorithms. Traditional approaches have drawn on heuristic and greedy strategies, while more recent methods incorporate adaptive and evolutionary techniques to navigate complex search spaces. Dynamic environments, such as cloud-based warehouses, further complicate selection through variable workloads and elastic resource allocation. Advances in workload prediction and real-time monitoring are fostering intelligent schemes that adjust materialisation sets in response to shifting usage patterns. The global significance of this research lies in its capacity to enhance business intelligence, reduce operational expenditure and support scalable analytics in sectors ranging from finance to scientific research. Concrete implementations in distributed systems and modern platforms demonstrate practical gains, reaffirming the enduring relevance of materialized view selection as a foundational component of high-performance data warehousing.

Research from Nature Portfolio

No recent Nature Portfolio content available.

Research from all publishers

Recent studies have showcased adaptive genetic algorithms that integrate query, maintenance and storage costs into a unified fitness function, yielding near-optimal view sets in distributed warehouse platforms such as Apache Hive. Simulation and real-world experiments confirm that adjustable crossover and mutation parameters prevent premature convergence while achieving significant query speed-ups. Another line of inquiry employs machine-learning-based cost estimators to predict query workloads, enabling dynamic view selection that adapts materialisation choices as usage evolves. These predictive models reduce manual tuning and improve responsiveness to workload fluctuations. A third strand of work addresses cloud-native warehouses, where storage tiering and auto-scaling features are incorporated into selection criteria. By modelling on-demand storage costs and elastic compute pricing, these approaches optimise total cost of ownership while maintaining high availability and performance under variable query loads.

Materialized View Selection in Data Warehousing publication trend

The graph below shows the total number of articles in materialized view selection in data warehousing across all publications each year (not limited to Nature Index journals).

Technical terms

Materialized view: A precomputed summary table derived from base data to accelerate query processing.

Data warehouse: A centralized repository that integrates and stores large volumes of historical and transactional data for analysis.

Cost model: A mathematical framework for estimating resource utilisation, including query execution, storage and maintenance overheads.

Workload: The collection of queries and data operations executed against a warehouse, characterised by frequency and complexity.

Adaptive algorithm: An optimisation technique that dynamically adjusts its parameters or strategies based on intermediate results or environmental changes.

References

1. Materialized View Selection Based on Adaptive Genetic Algorithm and Its Implementation with Apache Hive. International Journal of Computational Intelligence Systems (2015).