Parallel Document Stream Search for Twig Pattern Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
As document size increases, the memory usage and search time for tree representations grow exponentially, making it inefficient to search for twig patterns in large documents.
Innovation Solution
The document tree is simplified into multiple linear streams, allowing each processing thread to search for matching twig patterns in parallel by dividing the tree into disjoint sub-trees and searching sub-streams associated with each sub-tree.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a tree representation is used to store and search documents, then search accuracy for twig patterns is maintained, but memory usage grows exponentially as document size increases
Solution Approach 1:
The patent divides the document tree into multiple linear streams by separating node data from tree structure. Each stream contains nodes with the same label in preorder traversal sequence, allowing the large tree to be segmented into manageable linear sequences that reduce memory requirements while maintaining search capability through the stream-based twig pattern matching algorithm.
2Measurement precision
If a tree representation is used to store and search documents, then search accuracy is maintained, but search time increases exponentially as document size increases
Solution Approach 1:
The tree is segmented into linear streams, enabling the search algorithm to process nodes in a more efficient linear manner rather than traversing the entire tree structure. This segmentation reduces the exponential time complexity by allowing the algorithm to match twig patterns across streams using a linear-time approach.
Solution Approach 2:
The patent transforms the two-dimensional tree structure into a one-dimensional linear stream representation. This dimensional change simplifies the search problem from complex tree traversal to linear sequence processing, significantly reducing search time while maintaining the ability to match twig patterns through the linear arrangement of nodes.
3Quantity of substance
If the document tree is simplified into linear streams, then memory usage is reduced, but the complexity of handling tree relationships is increased
Solution Approach 1:
The patent performs preliminary action by pre-traversing the document tree in preorder sequence and organizing nodes into streams based on their labels before the actual search operation. This preprocessing step captures the essential tree relationship information (parent-child and ancestor-descendant relationships) in the linear stream structure, so that during search, the algorithm only needs to process the simplified stream data rather than re-analyzing complex tree structures.
4Productivity
If multiple processing threads search streams in parallel, then search speed increases, but coordination complexity between threads is increased
Solution Approach 1:
The patent segments the document streams into distinct portions that can be independently processed by multiple threads. Each thread is assigned specific streams or stream ranges to search for twig patterns, eliminating the need for complex thread coordination while enabling parallel processing. The segmentation is designed so that threads operate independently on disjoint data, maximizing search speed without introducing coordination overhead.
Data Source
AI summary
A system and method for searching a document for a query pattern. A plurality of streams may be stored each including a linear sequence of nodes. Each stream may be associated with nodes having a common label in a data tree of the document. A query pattern may be searched for in the streams by executing a plurality of threads. Each of two or more of the threads may be used to search different sub-streams of the plurality of streams. Each of the different sub-streams searched for by each thread in each stream may be uniquely correlated with one or more disjoint sub-trees of a partition of the tree into a plurality of sub-trees. The two or more of the plurality of threads may be executed in parallel. A result of the query pattern search may be generated using at least one of the threads.