Scenario

An SQLite database contains e-commerce data that needs to be exported to Parquet format for analytics processing. The export must preserve table relationships and provide metadata about the conversion.

Task

Write a Python script at /home/interview/export_to_parquet.py that connects to /home/interview/ecommerce.db, extracts all tables, converts each to Parquet format with Snappy compression, and creates a manifest file. Save outputs to /home/interview/parquet_export/ directory.

Note: pandas, pyarrow, and sqlite3 are already installed.

Database Schema

Manifest Requirements

Create manifest.json containing: export timestamp, compression type, and for each table: table name, file name, row count, file size, and foreign key relationships.

Example Manifest

{
  "export_timestamp": "2026-02-13T10:30:00",
  "compression": "snappy",
  "tables": [
    {
      "table_name": "orders",
      "file_name": "orders.parquet",
      "row_count": 200,
      "file_size_bytes": 23456,
      "foreign_keys": [
        {"column": "customer_id", "references_table": "customers", "references_column": "id"}
      ]
    }
  ]
}