Structured Document Search Indexing for Speed and Completeness
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing structured-document databases face inefficiencies in search operations due to the need for costly data scans and index type mixing, leading to decreased search speed when handling multiple index types and schema variations.
Innovation Solution
The proposed solution involves an apparatus and method that generates a hierarchical search condition, relieves structure and value constraints, and delays costly processes until data acquisition, using a structured-document searching apparatus with units for storing and indexing structured documents, structure information, and vocabulary indices to optimize search planning and execution.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If data scan is performed to verify structure constraints in structured documents, then search completeness is improved, but search speed deteriorates
Solution Approach 1:
The patent applies preliminary action by pre-creating structure indexes during document storage that capture hierarchical relationships and structural constraints. When a search query arrives, the system uses these pre-built indexes to quickly identify candidate documents that satisfy structure constraints, avoiding the need to perform costly data scans on all documents. This preliminary preparation enables fast structure verification while maintaining complete search coverage.
2Adaptability or versatility
If multiple index types are mixed to handle various schemas, then search versatility is improved, but search speed deteriorates
Solution Approach 1:
The patent implements universality by creating a unified structure index that can handle multiple document schemas and hierarchical structures through a common indexing framework. The structure index uses standardized representations of parent-child relationships, sibling relationships, and structural constraints that work across different document types and schemas. This universal approach eliminates the need to mix multiple specialized index types, thereby maintaining search versatility while improving search speed through a single optimized indexing mechanism.
3Manufacturing precision
If structure collating process is performed to verify hierarchical relationships, then structure constraint satisfaction is improved, but processing cost increases
Solution Approach 1:
The patent applies preliminary action by pre-computing and storing structural relationship information in the structure index during document ingestion. The index captures parent-child relationships, sibling relationships, and structural constraints in advance. During search operations, the system directly queries this pre-computed structure index to verify hierarchical relationships, eliminating the need to perform expensive structure collating processes on the actual document data. This maintains precise structure constraint satisfaction while dramatically reducing processing costs.
4Speed
If vocabulary index is attached to specific paths to speed up search, then search speed is improved, but index complexity increases
Solution Approach 1:
The patent implements universality by designing a structure index that serves multiple functions: it captures hierarchical relationships, stores structural constraints, enables structure-based queries, and supports vocabulary searches. Instead of creating separate vocabulary indexes for specific paths, the unified structure index integrates these capabilities through a common framework that represents document structures and constraints in a standardized way, thereby improving search speed without increasing index complexity.
Data Source
AI summary
A condition generating unit generates a hierarchical-type search condition including a search target structure ID and a search result structure ID. A first acquiring unit acquires an object ID corresponding to the search target structure ID to which a vocabulary index is not attached. A candidate generating unit generates a candidate of the search result in which an acquired object ID is associated with the search key as a first constraint condition. A second acquiring unit acquires a search result structure ID complying with a structure constraint. A result acquiring unit acquires an object corresponding to the object ID satisfying the first constraint condition.