Skip to content

README.md

Latest commit

 

History

History
83 lines (66 loc) · 4.01 KB

README.md

File metadata and controls

83 lines (66 loc) · 4.01 KB

Java to Python Migration Guide

Zero data migration required. CUBRID's JDBC driver and Python driver (pycubrid) use the same CAS broker protocol — your database, schema, and data stay exactly as-is.

Why Migrate?

Dimension Java (JDBC) Python (pycubrid) Python (SQLAlchemy)
Connection setup 8–12 lines 4 lines 3 lines
CRUD (full cycle) ~60 lines ~25 lines ~20 lines
Batch insert 1K rows ~30 lines 3 lines (executemany) 2 lines (add_all)
Transaction handling Manual try/catch/finally with context manager Session auto-rollback
Error handling Checked exceptions Python exceptions SQLAlchemy exceptions
ORM integration JPA/Hibernate (complex) SQLAlchemy (simple) Built-in

Lines of Code Comparison

                         Java JDBC    pycubrid    SQLAlchemy
Connection               12           4           3
Single INSERT            8            3           2
SELECT + iterate         15           5           4
Transaction block        20           6           5
Batch INSERT (1K rows)   30           3           2
─────────────────────────────────────────────────────────
Total                    85           21          16
Reduction                —            75%         81%

Performance

Our benchmark results show pycubrid performs well after recent optimizations:

  • SELECT 10K rows: 78ms total (fetch + parse)
  • Connection: 1.7ms
  • Single INSERT + COMMIT: ~55ms (COMMIT is the dominant cost at ~47ms)

COMMIT is 7× more expensive than INSERT execute. Batch your writes!

File Overview

File What You'll Learn
01_connection.py Connection patterns — JDBC DriverManager → pycubrid connect()
02_crud_dbapi.py CRUD with raw DB-API — PreparedStatement → parameterized queries
03_crud_orm.py CRUD with SQLAlchemy ORM — JPA Entity → Mapped class
04_transaction.py Transaction patterns — try/catch/finally → with blocks
05_batch_operations.py Batch operations — addBatch/executeBatch → executemany

Migration Checklist

  • Install Python dependencies: pip install -r requirements.txt
  • Replace JDBC connection strings with pycubrid connect() calls
  • Replace PreparedStatement with parameterized cursor.execute(sql, params)
  • Replace ResultSet iteration with fetchall() / fetchone() / fetchmany()
  • Replace try/catch/finally with Python with blocks or try/except/finally
  • Replace addBatch/executeBatch with cursor.executemany()
  • Replace JPA entities with SQLAlchemy models (optional — pycubrid works fine standalone)
  • Update CI/CD pipeline (Maven/Gradle → pip/Poetry)
  • Run existing SQL queries unchanged — CUBRID SQL syntax is identical regardless of driver

Key Differences to Watch

  1. Parameter markers: JDBC uses ? — pycubrid also uses ? (same!)
  2. Auto-commit: JDBC defaults to auto-commit ON — pycubrid defaults to auto-commit OFF
  3. Resource cleanup: JDBC requires explicit close() in finally — Python has with statements
  4. Null handling: JDBC wasNull() pattern — Python returns None naturally
  5. Type mapping: Java types map cleanly to Python types (see table below)

Type Mapping

CUBRID Type Java Type Python Type
INT int int
BIGINT long int
DOUBLE / FLOAT double float
VARCHAR / CHAR String str
DATE java.sql.Date datetime.date
TIME java.sql.Time datetime.time
TIMESTAMP java.sql.Timestamp datetime.datetime
BLOB / CLOB Blob / Clob bytes / str
NULL null + wasNull() None