Subsetting

Reduce records to create a smaller, representative subset of a relational database while maintaining referential integrity

Book a demo
subsetting data with Syntho
Health Plus Ai
KPN
Philips
Nederlandse Spoorwegen
Cedars Sinai
New compliance
SAS
Universiteit Leiden
Lifelines
Centraal Bureau voor de Statistiek
Euris Group
KVK
Leiden University Medical Center
UMCG

Key benefits of using
subsetting

Create smaller, representative datasets while preserving referential integrity.

Reduce infrastructure
and computational costs

Excessive data volumes can lead to high infrastructure and computation costs, which are unnecessary for test data in non-production environments. With subsetting capabilities, you can easily create smaller subsets of your data to reduce your costs.

Manageable test data by
testers and developers

Managing huge data volumes in non-production environments poses challenges for testers and developers. Smaller and thereby more manageable test data, significantly streamlining testing and development processes, ultimately optimizing the entire cycle in terms of time and resources.

Simplify test data management for faster setup and maintenance

Smaller data volumes facilitate faster and more straightforward setup and maintenance of non-production test environments. This is particularly relevant in complex IT landscapes and when frequent changes in data structures require regular updates and refreshes to ensure the representativeness of test data.

Enable secure testing, development, and training environments

By working with smaller, representative subsets of data, organizations can establish secure environments for testing, development, and training. This minimizes the risk of exposing sensitive information while maintaining data integrity and utility for non-production use cases.

User documentation

Explore the Syntho user documentation

Learn more

Subsetting in 3 steps

Subsetting in <span class="accent-for-white">3 steps</span>
01
Configure Table Settings

Include or Exclude tables for subsetting.

02
Adjust Rows to Generate

Define the row count in the Rows to generate field, where Synthesize creates rows using AI, Duplicate samples rows from the source, and Exclude skip generating rows; note that adjustments may impact foreign key relationships.

Product Demo

Subsetting

Create synthetic data that enhances the volume and diversity of your data

Coming soon: Advanced subsetting features

Subsetting is not as simple as
“just deleting data”

Subsetting is not as easy as simply deleting data, as all downstream and upstream related linked tables should be subsetting proportionally to preserve referential integrity.

Subsetting ensures that not only data in a target table is deleted, but also that any data in any other linked table related to the deleted data from the target table is deleted.

This ensures that referential integrity across tables, databases and systems is preserved as part of data deletion.

Reducing the data volume by removing “Person X” from “Table Y”, all records related to “Person X” in “Table Y” should be deleted, but also all records related to “Person X” in any other upstream or downstream related table (table A, B, C etc.) should also be deleted.

Reducing the data volume by removing “Richard” from the “Customers” table, all records related to “Richard” in the “Customer” table should be deleted, but also all records related to “Richard” in any other upstream or downstream related table (Payment table, Incidents table, Insurance Coverage Table etc.) should also be deleted.

Across tables
Across tables

Subsetting works across tables

Within databases
Within databases

Subsetting works within databases

Within systems
Within systems

Subsetting works within systems

Proportional subsetting

Coming soon

You can configure the Syntho Engine to subset a relational database and to ensure that all “linked tables” are subsetted based on the “Target Table”.

Proportional subsetting
Target table
Target table
These are all directly or indirectly connected tables to the “Target Table”. Links between tables may be direct, such as a target table listing allergies that reference a patient’s table through a foreign key relationship, or indirect, such as a target table referencing a patient’s table, which in turn references a hospital’s table.
Linked tables
Linked tables
These are all directly or indirectly connected tables to the “Target Table”. Links between tables may be direct, such as a target table listing allergies that reference a patient’s table through a foreign key relationship, or indirect, such as a target table referencing a patient’s table, which in turn references a hospital’s table.

Subsetting based on business rules

Coming soon

In addition to proportional subsetting, where you specify a percentage for data extraction, our advanced capabilities allow you to precisely define the target group for subsetting. For instance, you can specify criteria to include or exclude specific subsets, providing greater flexibility and control over the data extraction process

  • Customers younger than 60 years and older than 30 years and
  • As Male customers
Subsetting based on business rules

Frequently Asked Questions

What is subsetting?

Many organizations have production environments with massive amounts of data and do not want massive amounts of data in non-production test environments. Hence, database subsetting is used to create a smaller, representative subset of a larger relational database with preserved referential integrity. Organizations utilize sub-setting for test data to reduce costs, to make it manageable and for faster setup and maintenance.

What is referential integrity and why is it important?

Referential integrity is a concept in database management that ensures consistency and accuracy between tables in a relational database. Referential integrity would ensure that every value that corresponds to “Person 1” of “Table 1” corresponds to the correct value of “person 1” in “Table 2” and any other linked table.

Enforcing referential integrity is crucial for maintaining the reliability of test data in a relational database as part of non-production environments. It prevents data inconsistencies and ensures that relationships between tables are meaningful and reliable for proper testing and software development.

Test data in a relational database environment should preserve referential integrity to be usable.

View all FAQ’s

Build better and faster with synthetic data today

Unlock data access, accelerate development, and enhance data privacy.

Book a demoContact Us

Join our newsletter

Keep up to date with synthetic data news