Skip to main content
Loading navigation
Back to challenges

GovTech Cloud Legacy Migration & Retention Engine

Modernizing legacy public sector data storage requires automated mapping of legacy database schemas to cloud-native, compliant storage backends. Build a legacy migration translation engine that parses cold storage schemas and generates secure DuckDB migration scripts. Achieve 100% schema parity and regulatory retention flag preservation.

Workflow AutomationHosted by Vera
Sign in free to start

What you are building

Build an automated data schema migration and retention tag mapping engine for public sector cloud modernization.

How work is evaluated

Evaluation

Evaluates schema parity and target script execution correctness in DuckDB.

Datasets

Public datasets, plus the private datasets you have access to.

How submissions are scored

Dimensions
2 scoring checks
Binary
2 pass or fail dimensions
Ordinal
0 scaled dimensions
Dimension 1

duckdb_execution_check

Ensures generated SQL executes without error in DuckDB.

Binary check

This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.

Dimension 2

type_conversion_accuracy

Percentage of correctly mapped column data types. • target: 1 • range: 0-1

Binary check

This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.

Learning goals

  • Parse legacy DDL statements (Oracle/DB2 dialect) into normalized ASTs

  • Generate target DuckDB and Parquet schema definition scripts

  • Automate application of statutory data retention metadata tags

  • Validate data type parity and losslessness across migration steps

Resources

Dataset notes

Synthetic legacy database DDL definitions from legacy government administrative applications.

Challenge setup

Evaluates schema parity and target script execution correctness in DuckDB.

Challenge input

Legacy DDL string

DuckDB

Fast in-process SQL database.

Evaluated output

Target DuckDB SQL string

Evaluation checks
  • Ensures generated SQL executes without error in DuckDB.
  • Percentage of correctly mapped column data types. • target: 1 • range: 0-1
Configured evaluation evidence
  • Benchmark: DDL-Transpile-Bench
  • Type Conversion Accuracy target: 1
  • 1 public reference case
Configured runtime
  • Python execution harness
  • Python sandbox (unavailable on Versalist)

Shows the first task, up to three tools, and up to three evidence items.

View technical recipe

Configured tools

Action Space
  • Skyvern · Required
  • Weights & Biases · Optional
Reward / Eval
  • Distributional · Optional
Safety / Guardrails
  • Distributional · Optional

Evaluation contract

  • duckdb_execution_check · Weight 1
  • type_conversion_accuracy · Weight 1

Metrics

  • type_conversion_accuracy · Target 1

Benchmarks

  • DDL-Transpile-Bench
  • general

Reference cases

  • 1 public reference case

Workflow tasks

  1. Task 1

    Input: Legacy DDL string

    Output: Target DuckDB SQL string

    Execution harness: Python

Evaluation tools

  • DuckDB · Target Analytical Engine

Recipe freeze marker

No recipe freeze time is recorded.

Frequently Asked Questions about GovTech Cloud Legacy Migration & Retention Engine