Directed Graph Conversion for Key-Value to Relational Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Transaction-oriented systems with loosely-structured key-value databases face challenges in converting binary data to relational formats required for data mining and analytical processing, as existing methods struggle to handle diverse and complex data structures.
Innovation Solution
A system converts serialized object graphs in key-value databases to directed graphs, where each node represents an entity and edges represent relationships, using graph queries to identify unique entities and extract properties, thereby transforming the data into a relational format suitable for relational databases.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If loosely-structured binary data is stored in key-value databases, then data storage flexibility is improved, but data conversion to relational formats becomes difficult
Solution Approach 1:
The patent introduces an intermediary conversion system that includes a graph query processor and relational data generator. This intermediary layer translates between the flexible key-value store format and the structured relational format required for analytical processing, resolving the contradiction by providing a bridge that maintains both storage flexibility and queryability.
Solution Approach 2:
The conversion process is segmented into distinct stages: deserialization of binary data, construction of directed graphs from key-value pairs, application of graph queries to identify entities and relationships, and generation of relational tables. This segmentation makes the complex conversion process manageable and systematic.
2Productivity
If binary object graphs are converted to relational format, then data mining capability is improved, but data structure complexity increases
Solution Approach 1:
The patent extracts entities, properties, and relationships from the binary object graphs through graph queries, separating these elements into distinct relational tables. This extraction process enables data mining operations while organizing the complex data structure into manageable relational components that can be independently processed.
3Measurement precision
If graph queries are applied to identify unique entities, then data accuracy is improved, but processing time increases
Solution Approach 1:
The patent performs preliminary actions by pre-identifying entities and relationships through graph queries before the actual data conversion to relational format. This preliminary identification ensures accuracy in entity mapping while allowing the subsequent conversion process to proceed efficiently with already-analyzed data structures.
Data Source
AI summary
A key-value database may maintain a collection of data in which the value components are serialized object graphs. The object graphs may be deserialized and converted to a directed graph. Each node may represent an entity and each edge a relationship between entities. Instructions associated with a table may indicate criteria for identifying corresponding entities in a directed graph. A row of the table may be created for each entity having a unique permutation of identifying properties. Column values may be obtained using instructions for accessing entity properties and mapping from the properties to column values.