Synthetic Test Data Generation From Production Table Metadata

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing systems face challenges in generating test data for test systems due to high risks of re-identification of de-identified data, delays in data availability, and inability to perform performance testing with large volumes of test data, while also failing to mimic production data quality.

Innovation Solution

A system and method for synthetically generating test data using sample production data and rules, which mimics production data characteristics while protecting sensitive fields, thereby avoiding the need for de-identification algorithms and reducing processing resources.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If de-identification algorithms are applied to production data to generate test data, then data security is improved, but processing time and complexity increase

Engineering Contradiction:

Improvedata securityVSAvoidprocessing time

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The patent creates synthetic test data that copies the statistical properties, data types, and relationships of production data without copying actual sensitive values. A data generator uses metadata from production data to create artificial records that mimic the structure and characteristics of real data while containing no actual sensitive information, thus achieving both security and efficiency

Inventive Principle:

Principle #26Copying

Solution Approach 2:

The system transforms production data parameters into synthetic test data parameters by analyzing data properties (data types, length, format, relationships) and generating new values that match these parameters. This changes the actual data values while preserving the structural parameters, achieving de-identification without complex algorithms

Inventive Principle:

Principle #35Parameter changes

2Reliability

If de-identification algorithms are applied to production data to generate test data, then data security is improved, but processing complexity increases

Engineering Contradiction:

Improvedata securityVSAvoidprocessing complexity

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

Instead of applying complex de-identification algorithms to transform real data, the system copies only the essential characteristics (data types, formats, relationships) and generates new synthetic data from scratch. This simplifies processing by replacing complex transformation algorithms with straightforward data generation based on metadata templates

Inventive Principle:

Principle #26Copying

Solution Approach 2:

The system extracts only the necessary structural properties from production data (data types, length, format patterns, relationships) and leaves behind the actual sensitive values. This extraction approach simplifies processing by focusing only on what is needed for test data generation while automatically discarding sensitive information

Inventive Principle:

Principle #2Taking out (Extraction)

3Productivity

If production data is copied to test database, then test data availability is improved, but re-identification risk increases

Engineering Contradiction:

Improvetest data availabilityVSAvoidre-identification risk

Core Design Contradiction:

ProductivityVSObject-affected harmful factors

Solution Approach 1:

The system creates synthetic copies of production data that preserve the structure, relationships, and statistical properties necessary for testing, while containing no actual sensitive values. This copying approach provides immediate test data availability while inherently preventing re-identification since the data is artificially generated rather than copied from production

Inventive Principle:

Principle #26Copying

Solution Approach 2:

The patent introduces a data generator as an intermediary between production data and test data. This mediator analyzes production data metadata and generates synthetic test data without directly copying or transmitting actual sensitive values, thus eliminating re-identification risk while maintaining data utility for testing

Inventive Principle:

Principle #24Intermediary (Mediator)

4Productivity

If production data is copied to test database, then test data availability is improved, but processing time increases

Engineering Contradiction:

Improvetest data availabilityVSAvoiddata copying time

Core Design Contradiction:

ProductivityVSLoss of time

Solution Approach 1:

The system copies only the essential metadata (data types, formats, relationships) from production data and uses this information to rapidly generate synthetic test data. This approach is faster than copying actual production data because it avoids the time-consuming process of selecting, transforming, and de-identifying large volumes of real data

Inventive Principle:

Principle #26Copying

Data Source

PatentUS20260017179A1System and method for generating test data for a code testing system

Publication Date: 2026.01.15 BANK OF AMERICA CORP

AI summary

A data manager obtains table metadata associated with a production database table stored in a production database of the production system. In addition, the data manager extracts a portion of the production data from the production database table by running a query in the production database, wherein the extracted portion of the production data is to be used as sample data when part of generating the test data. The data manager determines data properties of the production data stored in the production database table based on the sample data extracted from the production database table. The data manager then generates a requested number of data records of the test data based on the table metadata and the data properties associated with the production database table, wherein the generated test data at least partially mimics the production data from the production database table.