Relational Database Tuning via SQL Pattern Aggregation

The patent groups multiple SQL commands that share similar search conditions and methods into acquisition patterns. By merging individual command analyses into pattern-based groupings, the system maintains precise tuning for each command type while reducing overall administrative workload. The generating unit creates tuning tables that apply optimized parameters across entire groups of similar commands rather than treating each command in isolation.

The patent creates acquisition patterns that serve as universal templates for multiple SQL commands. A single acquisition pattern can represent many different commands that share the same search characteristics, allowing the tuning system to handle diverse commands through a unified approach. This multi-functionality reduces complexity while maintaining the ability to precisely tune each command variant within the pattern.

The patent analyzes acquisition patterns to identify specific columns that benefit from indexing based on their actual usage in search conditions. Rather than automatically indexing all frequently accessed columns, the system selectively applies indexing only to columns where it provides local optimization value. The tuning table generates targeted index recommendations that improve search efficiency for specific pattern types without unnecessarily impacting update performance across the entire database.

The patent avoids excessive indexing by applying index generation selectively based on pattern analysis results. Instead of adding indexes to all columns that appear in SQL commands, the system identifies the subset of columns within each acquisition pattern that would most benefit from indexing. This partial action approach ensures search efficiency improvements while minimizing the negative impact on update operations.

The patent segments the large set of SQL commands into distinct acquisition patterns based on shared search conditions and methods. This segmentation divides the overwhelming task of analyzing all individual commands into manageable pattern groups. The extracting unit identifies commonalities among commands and groups them accordingly, allowing administrators to focus on optimizing patterns rather than individually analyzing each command in a large-scale database.

The patent extracts the essential characteristics (search conditions and search methods) from individual SQL commands to create acquisition patterns. By taking out and abstracting the common elements that define command behavior, the system reduces the complexity of bottleneck identification. Instead of examining every detail of each command, the extracting unit isolates the key pattern-defining features that determine optimization strategies.