Oracle SQL*Loader: Zero to Pro

A hands-on, runnable curriculum for learning Oracle SQL*Loader — from your very first CSV load to a production-style staging → validate → merge pipeline. Every level has real sample data, a real .ctl control file, and a README explaining exactly what's new and why it matters.

Built as a portfolio project to demonstrate Oracle database and ETL/data-loading skills.

Why this project exists

SQL*Loader is one of those tools that's everywhere in enterprise Oracle shops (data migrations, nightly batch feeds, vendor file ingestion) but rarely gets taught step by step. This repo is that missing tutorial — structured as a progression, not a wall of reference documentation.

Prerequisites

• Access to an Oracle database (Oracle XE is free and works fine — Oracle Database Free, or any Oracle instance/PDB you already have)
• Oracle Client tools installed, specifically sqlldr and sqlplus on your PATH
• A schema/user with CREATE TABLE privileges to practice in

Check your setup:

sqlldr help=y
sqlplus -v

Project structure

Project_SQL_Loader/
├── 01-Basics/                 Your first load: plain CSV, one table
├── 02-Intermediate/           Headers, quoted text, NULLs, load modes
├── 03-Advanced/
│   ├── multi_table/           One file loading into two tables
│   ├── fixed_width/           Positional (non-delimited) data
│   ├── when_clause/           Conditional loading, bad vs. discarded rows
│   └── direct_path/           Conventional vs. direct path performance
├── 04-Error-Handling/         Reading logs, deliberately messy data
├── 05-Real-World-Project/     Staging → validate → MERGE pipeline
└── docs/                      Cheat sheet + command-line reference

Learning path

Work through the folders in order — each one builds on the last:

#	Folder	You'll learn
1	01-Basics	Control file anatomy, APPEND, date masks, bad/discard files
2	02-Intermediate	Skipping headers, quoted/embedded commas, NULLIF, TRUNCATE
3	03-Advanced	Multi-table loads, fixed-width data, WHEN filtering, direct path
4	04-Error-Handling	Reading .log files fluently, ERRORS=, reconciling row counts
5	05-Real-World-Project	The staging-table pattern real pipelines use, plus a MERGE

Each folder has its own README.md with the exact commands to run and what to verify afterward. Start here:

Quick reference

See docs/sqlloader_cheat_sheet.md for a one-page syntax reference and docs/command_reference.md for command-line options — useful once you've done the levels once and just need a refresher.

Skills demonstrated

• Oracle SQL*Loader (conventional and direct path)
• Control file design: delimited, positional/fixed-width, and mixed formats
• Data quality handling: bad rows, discarded rows, NULL normalization
• ETL pattern: staging table → validation → MERGE into production
• Performance considerations: commit sizing, direct path, parallel loads
• Reading and reconciling sqlldr log output

License

MIT — see LICENSE. Use this freely for learning or as a template for your own portfolio project.