rpg-migration-analyzer（rpg 迁移分析器）

RPG Migration Analyzer

Analyzes legacy RPG programs (RPG III/IV/ILE) from AS/400 and IBM i systems for migration to modern Java applications, extracting business logic, data structures, file operations, and generating actionable migration strategies.

Overview

This skill provides comprehensive analysis and migration planning for RPG (Report Program Generator) applications. It extracts program specifications, converts RPG data types to Java equivalents, maps file operations to modern database access patterns, and generates implementation-ready Java code structures.

Key Migration Focus: RPG to Java with proper handling of packed decimals (BigDecimal), data structures (POJOs), file operations (JPA/JDBC), indicators (boolean variables), and business logic preservation.

When to Use This Skill

Use this skill when:

• Analyzing RPG source files (.rpg,.rpgle,.RPGLE) for modernization
• Planning migration from AS/400 or IBM i systems to Java
• Converting RPG data structures (D-specs) to Java classes
• Mapping RPG file operations (F-specs) to database access patterns
• Understanding RPG program dependencies and call chains
• Generating Java code equivalents from RPG business logic
• Estimating complexity and effort for RPG migration projects
• Creating migration documentation and strategy reports
• Modernizing legacy mainframe applications to microservices
• User mentions: RPG analysis, AS/400 migration, IBM i modernization, Report Program Generator, packed decimal conversion

Core Capabilities

1. Program Analysis

Extract and analyze RPG program components:

• Specification types: H-spec (header/control), F-spec (file definitions), D-spec (data definitions), C-spec (calculation/logic), P-spec (procedures)
• Data structures: D-specs with nested structures, arrays (DIM), external references (EXTNAME), qualifiers (LIKEDS, QUALIFIED)
• File definitions: Physical files, logical files, display files (WORKSTN), printer files
• Business logic: Calculation specifications, control structures (IF/ELSE/DO/FOR), expressions (EVAL)
• Indicators: Legacy indicators (*IN01-*IN99), built-in indicators (*INLR,*INOF)
• Built-in functions: String functions (%SUBST, %TRIM, %SCAN), date functions (%DATE, %DAYS), math functions (%DEC, %INT), file status (%EOF, %FOUND, %ERROR)
• Error handling: %ERROR, %STATUS, ON-ERROR blocks

2. Data Structure Mapping

Convert RPG data definitions to Java equivalents:

• D-spec conversion: Data structure definitions to Java classes (POJOs)
• Data type mapping:

- Packed decimal (P) → BigDecimal (preserve precision) - Zoned decimal (S) → BigDecimal (decimal with sign) - Character (A) → String - Date (D) → LocalDate - Time (T) → LocalTime - Timestamp (Z) → LocalDateTime - Indicator (N) → boolean - Binary integer (I) → int or long

• Arrays (DIM): Convert to Java List<T> or arrays T[]
• Nested data structures: Convert LIKEDS to nested Java classes
• External data structures (EXTNAME): Generate JPA entities from database table definitions
• Initialization (INZ): Map to Java field initializers or constructors

3. File Operations

Parse and convert RPG file I/O to modern database access:

• File types: Physical files (DISK), logical files (keyed access), display files (WORKSTN), printer files (PRINTER)
• Access methods: Sequential (full read), keyed (direct access by key), arrival sequence
• I/O operations:

- READ/READE → JPA query methods, JDBC ResultSet iteration - WRITE → JPA persist(), JDBC INSERT - UPDATE → JPA merge(), JDBC UPDATE - DELETE → JPA remove(), JDBC DELETE - CHAIN → JPA findById(), Optional pattern - SETLL/READE loop → JPA findBy...() queries with ordering

• File status: %EOF (end of file), %FOUND (record found), %ERROR (I/O error) → Java exceptions or Optional
• Transaction boundaries: Identify commit boundaries (COMMIT operation code)

4. Java Migration Strategy

Generate modern Java implementation patterns:

• POJOs: Plain Old Java Objects from D-spec data structures
• JPA Entities: @Entity annotations for database tables (from EXTNAME files)
• Repository pattern: Spring Data JPA repositories for file operations
• Service methods: Business logic from procedures and subroutines
• Bean Validation: @NotNull, @Size, @DecimalMin/Max from RPG field validations
• Exception handling: Convert %ERROR patterns to try-catch blocks with custom exceptions
• Collections: Java Collections API (List, Map, Set) from RPG arrays and data structures
• DTOs: Data Transfer Objects for service boundaries
• Transaction management: @Transactional annotations for commit boundaries

5. Dependency Analysis

Map program relationships and external dependencies:

• Program calls: CALLB (bound procedure calls), CALLP (prototyped procedure calls)
• Service programs: BNDDIR (binding directories), *SRVPGM objects
• File dependencies: All physical/logical files accessed by the program
• Database tables: DB2 for i tables referenced (EXTNAME)
• /COPY members: Include files, copy source members, prototypes
• Call chains: Identify calling programs and called programs
• Shared data areas: *DTAARA usage
• Message queues: QMHSNDPM (send program message)

Instructions

Follow these steps to analyze and migrate RPG programs to Java:

Step 1: Locate RPG Source Files

Find RPG source files (.rpg,.rpgle,.RPGLE extensions for RPG III/IV/ILE free-format).

find . -name "*.rpg" -o -name "*.rpgle" -o -name "*.RPGLE"

Step 2: Analyze Program Structure

Extract specifications (H, F, D, C, P), data structures, file definitions, procedures, and dependencies.

Automation: Run scripts/extract-structure.py for automated extraction.

Step 3: Map Data Types

Convert RPG to Java types - CRITICAL: Always use BigDecimal for packed/zoned decimals (never float/double).

Step 4: Convert Code Patterns

Transform RPG operations to Java - key conversions:

• Calculations: EVAL expressions → BigDecimal arithmetic methods
• File I/O: CHAIN → findById() with Optional, READ → query methods
• Arrays: Adjust 1-based (RPG) to 0-based (Java) indexing
• Strings: %SUBST(1-based) → substring(0-based)
• Indicators: *IN01 → named boolean variables

See pseudocode-rpg-rules.md for comprehensive conversion patterns.

Step 5: Generate Java Implementation

Create:

1. POJOs from D-spec data structures (scripts/generate-java-classes.py)
2. JPA entities for database tables
3. Repository interfaces (Spring Data JPA)
4. Service methods for business logic
5. Exception handling and validation

Step 6: Analyze Dependencies

Map program calls (CALLB/CALLP), file dependencies, /COPY members, service programs.

Automation: Run scripts/analyze-dependencies.sh or .ps1

Step 7: Create Migration Report

Generate documentation with program overview, dependencies, data mappings, Java design, and complexity estimate.

Template: Use assets/migration-report-template.md

Step 8: Validate and Test

Verify: BigDecimal usage, index adjustments, transaction boundaries, error handling, unit tests with AS/400 data samples.

Quick Reference

Critical Migration Rules

1. ALWAYS use BigDecimal for RPG packed (P) and zoned (S) decimals - never float/double
2. Adjust indexing: RPG uses 1-based arrays/strings, Java uses 0-based
3. Replace indicators: Convert *IN01-*IN99 to descriptive boolean variables
4. File operations: CHAIN → findById(), READ → query methods with Optional
5. String functions: %SUBST(1:10) → substring(0, 10) - adjust positions
6. Date operations: RPG date functions → LocalDate/LocalTime API
7. Transactions: Identify COMMIT operations → @Transactional annotations
8. Error handling: %ERROR/%STATUS → try-catch with custom exceptions

Example: Data Structure to Java Class

RPG D-spec:

D Employee   DS
D   EmpId             6  0
D   EmpName          30  A
D   Salary           63  2P

Java POJO:

public class Employee {
    private int empId;
    private String empName;
    private BigDecimal salary;  // 6 digits, 2 decimals
    // getters/setters
}

Example: File Operation Conversion

RPG CHAIN:

C     custId  CHAIN  CUSTFILE
C             IF     %FOUND(CUSTFILE)

Java with JPA:

customerRepository.findById(custId).ifPresent(customer -> {
    // process customer
});

// Service usage
public class CustomerService {
    @Autowired
    private CustomerRepository customerRepository;

    public Optional<Customer> findCustomer(Integer custId) {
        return customerRepository.findById(custId);
    }
}

Edge Cases

Case 1: Packed Decimal Precision

Problem: Using double/float causes precision errors. Solution: Always use BigDecimal from String literals: new BigDecimal("123.45")

Case 2: Array Index Shift

Problem: RPG 1-based, Java 0-based. Solution: Adjust all array/string index references. Test thoroughly.

Case 3: External Data Structures

Problem: EXTNAME without DDL source. Solution: Use DSPFFD command, query DB2 SYSTABLES/SYSCOLUMNS, or create entities from runtime data.

Case 4: Legacy Indicators

Problem: *IN01-*IN99 for control flow. Solution: Replace with descriptive booleans: boolean invalidAmount = false;

Case 5: Date Century Handling

Problem: 2-digit years (Y2K). Solution: Use 4-digit LocalDate, apply century window logic, document assumptions.

Guidelines

1. BigDecimal mandatory: Never float/double for packed/zoned decimals
2. Named booleans: Replace *IN01-99 with descriptive names
3. Database access: Map file I/O to JPA/JDBC operations
4. JPA entities: Create from EXTNAME physical file definitions
5. Exception handling: Convert %ERROR/%FOUND to exceptions/Optional
6. Test with AS/400 data: Validate with actual legacy system data
7. Transactions: Identify COMMIT operations → @Transactional
8. Document rules: Extract and document implicit business logic
9. Character encoding: Verify EBCDIC → Unicode conversions
10. Batch processing: Convert batch jobs to Spring Batch framework

Error Handling

Type 1: File I/O Errors

Detection: %ERROR or %STATUS checks. Handling: Use try-catch with custom exceptions (CustomerNotFoundException, DataAccessException)

Type 2: Arithmetic Overflow

Detection: Insufficient field size. Handling: BigDecimal with appropriate scale/precision, catch ArithmeticException

Type 3: Missing Dependencies

Detection: Missing /COPY members. Handling: Track all includes, create shared interfaces, use Maven/Gradle dependencies

Additional Resources

See detailed documentation in the references/ directory:

RPG Translation Rules (Organized by Topic)

The RPG translation rules are organized into focused, topic-specific files for easier navigation:

• pseudocode-rpg-rules.md - Master index with quick start guide and file navigation
• pseudocode-rpg-core-rules.md - Foundation: specs, data types, basic operations, file I/O
• pseudocode-rpg-functions.md - Built-in functions (BIFs): string, date/time, math, array functions
• pseudocode-rpg-data-structures.md - Data structure patterns: QUALIFIED, LIKEDS, OVERLAY, I/O specs
• pseudocode-rpg-patterns.md - Common idioms, translation patterns, pitfalls, critical rules
• pseudocode-rpg-advanced.md - ILE RPG, embedded SQL, web services, IFS, XML/JSON, threading
• pseudocode-rpg-migration-guide.md - Migration workflow, best practices, refactoring strategies

Other Reference Documentation

• pseudocode-common-rules.md - General pseudocode syntax and conventions
• testing-strategy.md - Testing approach for RPG to Java migration validation
• transaction-handling.md - AS/400 transaction patterns to Java transaction management
• performance-patterns.md - Performance optimization patterns for migrated code
• messaging-integration.md - Message queue and integration patterns for IBM i systems

Scripts

Python and shell scripts for automated analysis in scripts/:

• analyze-dependencies.sh/ps1 - Scans RPG source for CALLB, CALLP, /COPY; generates dependency graph
• extract-structure.py - Parses RPG specs (H, F, D, C, P); outputs structured JSON
• generate-java-classes.py - Creates Java POJOs from RPG data structures with proper types
• estimate-complexity.py - Calculates migration complexity score and effort estimate

Templates

• migration-report-template.md - Standard format for migration analysis reports
• java-class-template.java - Template for generated Java classes

analyze-dependencies.sh /.ps1

• Scans RPG source files for CALLB, CALLP, /COPY references
• Generates dependency graph in JSON format
• Identifies circular dependencies

Usage:

./scripts/analyze-dependencies.sh /path/to/rpg/source

extract-structure.py

• Extracts program structure (H/F/D/C/P specs)
• Lists all variables, data structures, files
• Identifies subroutines and procedures
• Outputs JSON structure file

Usage:

python scripts/extract-structure.py PROGRAM.rpgle --output structure.json

generate-java-classes.py

• Generates Java POJO classes from RPG data structures
• Creates proper field types (BigDecimal for packed decimals)
• Adds getters, setters, constructors
• Generates Bean Validation annotations

Usage:

python scripts/generate-java-classes.py structure.json --output-dir ./src/main/java

estimate-complexity.py

• Calculates migration complexity score
• Analyzes LOC, dependencies, file operations
• Provides effort estimate (hours/days)
• Generates priority ranking

Usage:

python scripts/estimate-complexity.py structure.json --report complexity-report.md

Templates

Use the migration report template for consistent documentation:

• migration-report-template.md - Standard format for migration analysis reports
• java-class-template.java - Template for generated Java classes with proper structure

Integration with Development Tools

This skill integrates with various development and analysis tools:

IBM i / AS/400 Tools

• Source Entry Utility (SEU): Extract source code from AS/400
• Programming Development Manager (PDM): Access member lists and source files
• WRKMBRPDM: Work with source members
• DSPFFD: Display file field descriptions for database structure analysis
• DSPPGMREF: Display program references and dependencies

Database Tools

• DB2 for i: Query system catalogs (SYSTABLES, SYSCOLUMNS) for metadata
• IBM Data Studio: Visual database design and SQL development
• DBeaver: Universal database tool with DB2 support

Modern Development Environment

• IntelliJ IDEA: Java development with Spring Boot support
• Eclipse: Java IDE with JPA tooling
• VS Code: Lightweight editor with Java extensions
• Git: Version control for both legacy source and new Java code

Migration Support Tools

• Spring Initializr: Bootstrap Spring Boot projects
• JPA Buddy: IntelliJ plugin for JPA entity generation
• Liquibase/Flyway: Database migration version control
• Maven/Gradle: Build automation and dependency management

Testing and Validation

• JUnit 5: Unit testing framework
• Spring Boot Test: Integration testing support
• Mockito: Mocking framework for unit tests
• TestContainers: Database integration testing with containers