Relational data extraction: Extract™

A production database that fits on your laptop

Take a compact, referentially complete slice of production: pick the tables, set the row counts, and Xcase follows every relationship.
Every parent, child, and lookup row the slice depends on is included - and nothing that it doesn't need.
Dramatic volume reduction with every foreign key intact, ready for development, test, and diagnosis.

Download a free trial

Includes Xcase Core. Synthetic data generation and datasets/snapshots are part of Core; Extract adds relationship-aware extraction from production and reference databases.

Extraction sets with per-table row counts, conditions, and CSV-driven selection

Define specific criteria such as random samples, targeted or conditional data, and enable automated recursion to include all relevant related data. Set rules at the relationship level to incorporate associated data recursively. You can even configure it to extract children for only a percentage of a parent's records. This approach maintains the representativeness of the extracted data while keeping it concise.
The automated dataset curation ensures consistency and accuracy by adhering to relational rules defined at the model level, even when those relationships are not enabled in the database.

Documentation

Relationship-aware inclusion of parent and child records

Pick the tables you need and Xcase follows the relationships: every parent, child and lookup row the slice depends on is included - and nothing it doesn't need. Instead of a full production copy you get a concise, representative subset that keeps referential integrity and stays focused on the business area you are testing.

The subset joins, because it was built from the relationships.
Focused on relevant business areas, small enough for a laptop.

Related diagram of the full extraction scope

Before anything runs, Xcase draws the extraction scope as a diagram: the tables you picked and every table the relationships pull in. The whole footprint of the slice, visible at a glance - including what stays out.

Volume reporting: source vs target size per table

For every table in scope, the volume report puts the source size and the extracted size side by side. Verify the reduction before you run, and show afterwards that the slice really is laptop-sized - with every foreign key intact.

Generated, reviewable extraction scripts

The extraction is emitted as scripts you can read: see exactly what will be selected and how, keep the scripts under version control, and rerun them whenever the non-production database needs a fresh slice.

Refresh non-production databases from extraction sets

The extracted dataset updates the non-production database: replace it entirely, or modify it in place while preserving its unique content. A repeatable refresh keeps development and test close to production - without ever copying production.

See Extract in action