Selective Multidimensional Index Scanning via Z-Order Curves

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing data management systems face inefficiencies in querying large datasets due to the need to scan entire indexes, which wastes resources and time, especially when only a portion of the index contains relevant data.

Innovation Solution

Implementing a multidimensional index that selectively scans only relevant portions of the index by using space filling curves like Z-order curves to map dimensions into a single dimension, allowing for incremental and targeted scanning based on query predicates, thereby skipping irrelevant parts.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If the entire index is scanned to process queries, then all possible matching data can be found, but I/O bandwidth and processing time are wasted on irrelevant portions

Engineering Contradiction:
Improvequery result completenessVSAvoidquery processing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent divides the multidimensional index into multiple partitions or segments that can be independently evaluated. By segmenting the index based on spatial regions or data blocks, the system can selectively scan only those segments that may contain relevant data for a given query, rather than scanning the entire index. This reduces query processing time while maintaining result completeness.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent implements partial scanning of the index by using bounding box calculations and spatial filters to identify and scan only the necessary portions of the index that could potentially contain matching data. This partial action approach avoids the excessive scanning of irrelevant index portions, significantly reducing I/O operations and processing time while still ensuring all matching records are found.

Inventive Principle:
Principle #16Partial or excessive action

2Measurement precision

If the entire index is scanned to ensure complete query results, then accuracy is maintained, but I/O bandwidth is wasted

Engineering Contradiction:
Improvequery result accuracyVSAvoidI/O bandwidth
Core Design Contradiction:
Measurement precisionVSLoss of energy

Solution Approach 1:

The patent performs preliminary actions by pre-calculating and storing metadata about index partitions, such as minimum and maximum bounding boxes for each spatial region. Before executing a query, the system uses this pre-computed metadata to quickly determine which index partitions could possibly contain matching data, eliminating the need to scan partitions that definitely don't match. This preliminary filtering maintains query accuracy while dramatically reducing I/O bandwidth consumption.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent extracts and utilizes spatial metadata (such as bounding box coordinates, partition boundaries, and data distribution statistics) from the multidimensional index structure. By extracting this summary information, the system can make intelligent decisions about which index portions to scan without reading the actual data, thereby maintaining result accuracy while minimizing I/O operations.

Inventive Principle:
Principle #2Taking out (Extraction)

3Productivity

If more dimensions are added to the index to improve querying capabilities, then query optimization increases, but index complexity increases

Engineering Contradiction:
Improvequery processing efficiencyVSAvoidindex structure complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent handles multidimensional data by transforming it into a different dimensional representation suitable for indexing. Rather than directly indexing complex multidimensional relationships, the system projects or transforms the multidimensional data into a form that can be efficiently organized and queried, balancing the benefit of multidimensional querying with manageable index complexity.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

Data Source

PatentUS11030177B1Selectively scanning portions of a multidimensional index for processing queries
Publication Date: 2021.06.08 AMAZON TECH INC
  • US11030177B1 patent drawing
  • US11030177B1 patent drawing
  • US11030177B1 patent drawing

AI summary

Portions of a multidimensional index for a database table may be selectively scanned for processing queries. A query may be received for a database table with a multidimensional index. A range of the multidimensional index may be identified for processing the query. Items mapped to different portions of the query may be scanned to apply the query. Some portions adjacent to a scanned portion may be skipped upon a determination that the adjacent portion does not include items that can satisfy the query. A result based on the scan operations can be provided in response to the query.