Context-Aware Search Indexing for Unstructured XML Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current database systems fail to perform context-aware searches on XML data stored as raw text in unstructured columns, as they cannot leverage the tagged content and hierarchical structure of XML documents, leading to loss of document fidelity and inefficient querying.
Innovation Solution
A database server indexes marked-up documents in unstructured columns using value-to-node and node-to-physical-location mappings, allowing context-aware searches without decomposing the documents into relational parts, and supports both XQuery-based full-text search for XML data and plain text search for non-XML data within the same framework.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If XML data is stored as raw text in unstructured columns, then storage flexibility and user convenience are improved, but the ability to perform context-aware searches is lost
Solution Approach 1:
The patent introduces an intermediary indexing layer that sits between the raw XML text storage and the search operation. The index stores pre-parsed contextual information (element names, hierarchies, attributes) without requiring the XML data to be stored in structured columns. This intermediary structure enables context-aware searches while preserving the flexibility of storing XML as raw text in unstructured CLOB columns.
Solution Approach 2:
The patent performs preliminary parsing and indexing of XML contextual structure during data insertion or in background operations, before search operations are executed. The index is pre-computed with element hierarchies, names, and relationships, so that when a search is performed, the system can quickly navigate the pre-built index structure rather than parsing the entire XML document at query time.
2Difficulty of detecting and measuring
If XML data is decomposed into relational parts for structured storage, then context-aware search capability is improved, but document fidelity is lost
Solution Approach 1:
The patent segments the indexing function from the storage function. The XML document itself is stored intact as a single unit in unstructured columns, preserving document fidelity. Separately, the indexing structure segments the contextual information (element names, hierarchies, attributes) into discrete indexable units that enable context-aware searches without altering or decomposing the original document.
Solution Approach 2:
The patent creates a copy of the contextual structure information from the XML document for indexing purposes, while the original XML document remains unchanged in its raw text form. The index is a separate representation that captures the hierarchical and contextual relationships without modifying or decomposing the source document, thus preserving document fidelity while enabling efficient searches.
3Difficulty of detecting and measuring
If XML data is stored in structured XMLType columns, then context-aware search is enabled, but storage flexibility and user convenience are reduced
Solution Approach 1:
The patent creates a universal indexing framework that works with XML data stored in unstructured columns, eliminating the need for separate structured XMLType column handling. The same indexing mechanisms and search operations that would work on structured XML data now work on raw text XML storage, providing multi-functionality across different storage approaches and maintaining storage flexibility while enabling context-aware searches.
4Quantity of substance
If full-text search is performed on raw XML text, then search coverage is improved, but search precision and context awareness are lost
Solution Approach 1:
The patent applies local quality by creating an index that preserves the contextual quality of different parts of the XML document. Instead of treating all text uniformly, the index structure captures element-specific contexts, hierarchies, and attributes, allowing searches to be performed with both broad coverage and high precision by leveraging the contextual information stored in the index for each location in the document.
Data Source
AI summary
A method, computing device, and a non-transitory computer-readable medium is provided for facilitating a context-aware search on documents stored in an unstructured database column. The database server determines whether a document stored in an unstructured database object is marked up according to a recognized markup language. Documents that conform to the markup language are indexed with value-to-node mappings and node-to-physical-location mappings to provide a context-aware search of marked up nodes within the documents. Documents that do not conform to the markup language are indexed with value-to-document mappings to provide a full-text search of the documents. The database server evaluates a context-aware search against documents that conform to the markup language within the unstructured column without requiring the user identify the documents as XML documents and store the documents in an XMLType structured column.