What's in the download

The files in your dataset zip and how each database loads them.

datagenstudio-postgres-18-csv.zip
├── data/
│   ├── customers.csv          ← one file per table, in your chosen format
│   ├── sales.csv
│   └── support_tickets.csv
├── create_tables.sql          ← CREATE TABLE statements, parents first, with keys and links
├── load_files.sql             ← database-specific loader for the data files (when available)
├── insert_backup.sql          ← plain-SQL INSERTs — works without the data files
└── README.md                  ← numbered loading steps for your database

How loading works by database

Database	CSV	JSON / Parquet
PostgreSQL	`psql` with `\copy`	DuckDB (`load_files.duckdb.sql`)
MySQL / MariaDB	`LOAD DATA LOCAL INFILE`	DuckDB
SQLite	`sqlite3` shell (3.41+)	DuckDB
SQL Server	`BULK INSERT` via `sqlcmd`	`insert_backup.sql`
Oracle	`insert_backup.sql` (SQLcl / SQL*Plus)	`insert_backup.sql`
Snowflake	`PUT` + `COPY INTO` (Snowflake CLI)	`PUT` + `COPY INTO`
Azure SQL / other	`insert_backup.sql`	`insert_backup.sql`

Notes

PDF columns are hex-encoded in CSV/JSON files and stored as raw bytes in Parquet.
If a column’s values don’t fit its database type — a patterned ID like CUST-000001, Y/N flags, or dates in a custom format like 03/14/2024 — it’s stored as text to keep your format. The zip’s README lists these columns.
insert_backup.sql is always the most portable option: run it after create_tables.sql with any SQL client.