motovaultpro/data/vehicle-etl/migrations/001_create_vehicle_database.sql

-- Migration: Create Automotive Vehicle Selection Database
-- Optimized for dropdown cascade queries
-- Date: 2025-11-10

-- Drop existing tables if they exist
DROP TABLE IF EXISTS vehicle_options CASCADE;
DROP TABLE IF EXISTS engines CASCADE;
DROP TABLE IF EXISTS transmissions CASCADE;
DROP INDEX IF EXISTS idx_vehicle_year;
DROP INDEX IF EXISTS idx_vehicle_make;
DROP INDEX IF EXISTS idx_vehicle_model;
DROP INDEX IF EXISTS idx_vehicle_trim;
DROP INDEX IF EXISTS idx_vehicle_composite;

-- Create engines table with detailed specifications
CREATE TABLE engines (
    id SERIAL PRIMARY KEY,
    name VARCHAR(255) NOT NULL,
    displacement VARCHAR(50),
    configuration VARCHAR(50),
    horsepower VARCHAR(100),
    torque VARCHAR(100),
    fuel_type VARCHAR(100),
    fuel_system VARCHAR(255),
    aspiration VARCHAR(100),
    specs_json JSONB,
    created_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP,
    updated_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP
);
-- Prevent duplicate engine display names (case-insensitive)
CREATE UNIQUE INDEX IF NOT EXISTS uq_engines_name_lower ON engines (LOWER(name));

CREATE INDEX idx_engines_displacement ON engines(displacement);
CREATE INDEX idx_engines_config ON engines(configuration);

-- Create transmissions table
CREATE TABLE transmissions (
    id SERIAL PRIMARY KEY,
    type VARCHAR(100) NOT NULL,
    speeds VARCHAR(50),
    drive_type VARCHAR(100),
    created_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP,
    updated_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP
);
-- Prevent duplicate transmission display names (case-insensitive)
CREATE UNIQUE INDEX IF NOT EXISTS uq_transmissions_type_lower ON transmissions (LOWER(type));

CREATE INDEX idx_transmissions_type ON transmissions(type);

-- Create denormalized vehicle_options table optimized for dropdown queries
CREATE TABLE vehicle_options (
    id SERIAL PRIMARY KEY,
    year INTEGER NOT NULL,
    make VARCHAR(100) NOT NULL,
    model VARCHAR(255) NOT NULL,
    trim VARCHAR(255) NOT NULL,
    engine_id INTEGER REFERENCES engines(id) ON DELETE SET NULL,
    transmission_id INTEGER REFERENCES transmissions(id) ON DELETE SET NULL,
    created_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP,
    updated_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP
);
-- Prevent duplicate vehicle option rows
CREATE UNIQUE INDEX IF NOT EXISTS uq_vehicle_options_full ON vehicle_options (
    year, make, model, trim, engine_id, transmission_id
);

-- Indexes for cascading dropdown performance
CREATE INDEX idx_vehicle_year ON vehicle_options(year);
CREATE INDEX idx_vehicle_make ON vehicle_options(make);
CREATE INDEX idx_vehicle_model ON vehicle_options(model);
CREATE INDEX idx_vehicle_trim ON vehicle_options(trim);
CREATE INDEX idx_vehicle_year_make ON vehicle_options(year, make);
CREATE INDEX idx_vehicle_year_make_model ON vehicle_options(year, make, model);
CREATE INDEX idx_vehicle_year_make_model_trim ON vehicle_options(year, make, model, trim);
CREATE INDEX idx_vehicle_year_make_model_trim_engine ON vehicle_options(year, make, model, trim, engine_id);
CREATE INDEX idx_vehicle_year_make_model_trim_trans ON vehicle_options(year, make, model, trim, transmission_id);

-- Full-text search index for admin catalog search
CREATE INDEX idx_vehicle_options_fts ON vehicle_options
USING gin(to_tsvector('english', year::text || ' ' || make || ' ' || model || ' ' || trim));

-- Index on engines.name for join performance during search
CREATE INDEX idx_engines_name ON engines(name);

-- Views for dropdown queries

-- View: Get all available years
CREATE OR REPLACE VIEW available_years AS
SELECT DISTINCT year
FROM vehicle_options
ORDER BY year DESC;

-- View: Get makes by year
CREATE OR REPLACE VIEW makes_by_year AS
SELECT DISTINCT year, make
FROM vehicle_options
ORDER BY year DESC, make ASC;

-- View: Get models by year and make
CREATE OR REPLACE VIEW models_by_year_make AS
SELECT DISTINCT year, make, model
FROM vehicle_options
ORDER BY year DESC, make ASC, model ASC;

-- View: Get trims by year, make, and model
CREATE OR REPLACE VIEW trims_by_year_make_model AS
SELECT DISTINCT year, make, model, trim
FROM vehicle_options
ORDER BY year DESC, make ASC, model ASC, trim ASC;

-- View: Get complete vehicle configurations with engine and transmission details
CREATE OR REPLACE VIEW complete_vehicle_configs AS
SELECT
    vo.id,
    vo.year,
    vo.make,
    vo.model,
    vo.trim,
    e.name AS engine_name,
    e.displacement,
    e.configuration,
    e.horsepower,
    e.torque,
    e.fuel_type,
    t.type AS transmission_type,
    t.speeds AS transmission_speeds,
    t.drive_type
FROM vehicle_options vo
LEFT JOIN engines e ON vo.engine_id = e.id
LEFT JOIN transmissions t ON vo.transmission_id = t.id
ORDER BY vo.year DESC, vo.make ASC, vo.model ASC, vo.trim ASC;

-- Function to get makes for a specific year
CREATE OR REPLACE FUNCTION get_makes_for_year(p_year INTEGER)
RETURNS TABLE(make VARCHAR) AS $$
BEGIN
    RETURN QUERY
    SELECT DISTINCT vehicle_options.make
    FROM vehicle_options
    WHERE vehicle_options.year = p_year
    ORDER BY vehicle_options.make ASC;
END;
$$ LANGUAGE plpgsql;

-- Function to get models for a specific year and make
CREATE OR REPLACE FUNCTION get_models_for_year_make(p_year INTEGER, p_make VARCHAR)
RETURNS TABLE(model VARCHAR) AS $$
BEGIN
    RETURN QUERY
    SELECT DISTINCT vehicle_options.model
    FROM vehicle_options
    WHERE vehicle_options.year = p_year
    AND vehicle_options.make = p_make
    ORDER BY vehicle_options.model ASC;
END;
$$ LANGUAGE plpgsql;

-- Function to get trims for a specific year, make, and model
CREATE OR REPLACE FUNCTION get_trims_for_year_make_model(p_year INTEGER, p_make VARCHAR, p_model VARCHAR)
RETURNS TABLE(trim_name VARCHAR) AS $$
BEGIN
    RETURN QUERY
    SELECT DISTINCT vehicle_options.trim
    FROM vehicle_options
    WHERE vehicle_options.year = p_year
    AND vehicle_options.make = p_make
    AND vehicle_options.model = p_model
    ORDER BY vehicle_options.trim ASC;
END;
$$ LANGUAGE plpgsql;

-- Function to get engine and transmission options for a specific vehicle
CREATE OR REPLACE FUNCTION get_options_for_vehicle(p_year INTEGER, p_make VARCHAR, p_model VARCHAR, p_trim VARCHAR)
RETURNS TABLE(
    engine_name VARCHAR,
    engine_displacement VARCHAR,
    engine_horsepower VARCHAR,
    transmission_type VARCHAR,
    transmission_speeds VARCHAR,
    drive_type VARCHAR
) AS $$
BEGIN
    RETURN QUERY
    SELECT
        e.name,
        e.displacement,
        e.horsepower,
        t.type,
        t.speeds,
        t.drive_type
    FROM vehicle_options vo
    LEFT JOIN engines e ON vo.engine_id = e.id
    LEFT JOIN transmissions t ON vo.transmission_id = t.id
    WHERE vo.year = p_year
    AND vo.make = p_make
    AND vo.model = p_model
    AND vo.trim = p_trim;
END;
$$ LANGUAGE plpgsql;

-- Helper functions for trim-level options and pair-safe filtering
CREATE OR REPLACE FUNCTION get_transmissions_for_vehicle(p_year INTEGER, p_make VARCHAR, p_model VARCHAR, p_trim VARCHAR)
RETURNS TABLE(
    transmission_id INTEGER,
    transmission_type VARCHAR
) AS $$
BEGIN
    RETURN QUERY
    SELECT DISTINCT
        t.id,
        t.type
    FROM vehicle_options vo
    JOIN transmissions t ON vo.transmission_id = t.id
    WHERE vo.year = p_year
    AND vo.make = p_make
    AND vo.model = p_model
    AND vo.trim = p_trim
    ORDER BY t.type ASC;
END;
$$ LANGUAGE plpgsql;

CREATE OR REPLACE FUNCTION get_engines_for_vehicle(p_year INTEGER, p_make VARCHAR, p_model VARCHAR, p_trim VARCHAR)
RETURNS TABLE(
    engine_id INTEGER,
    engine_name VARCHAR
) AS $$
BEGIN
    RETURN QUERY
    SELECT DISTINCT
        e.id,
        e.name
    FROM vehicle_options vo
    JOIN engines e ON vo.engine_id = e.id
    WHERE vo.year = p_year
    AND vo.make = p_make
    AND vo.model = p_model
    AND vo.trim = p_trim
    ORDER BY e.name ASC;
END;
$$ LANGUAGE plpgsql;

CREATE OR REPLACE FUNCTION get_transmissions_for_vehicle_engine(p_year INTEGER, p_make VARCHAR, p_model VARCHAR, p_trim VARCHAR, p_engine_name VARCHAR)
RETURNS TABLE(
    transmission_id INTEGER,
    transmission_type VARCHAR
) AS $$
BEGIN
    RETURN QUERY
    SELECT DISTINCT
        t.id,
        t.type
    FROM vehicle_options vo
    JOIN engines e ON vo.engine_id = e.id
    JOIN transmissions t ON vo.transmission_id = t.id
    WHERE vo.year = p_year
    AND vo.make = p_make
    AND vo.model = p_model
    AND vo.trim = p_trim
    AND e.name = p_engine_name
    ORDER BY t.type ASC;
END;
$$ LANGUAGE plpgsql;

CREATE OR REPLACE FUNCTION get_engines_for_vehicle_trans(p_year INTEGER, p_make VARCHAR, p_model VARCHAR, p_trim VARCHAR, p_trans_type VARCHAR)
RETURNS TABLE(
    engine_id INTEGER,
    engine_name VARCHAR
) AS $$
BEGIN
    RETURN QUERY
    SELECT DISTINCT
        e.id,
        e.name
    FROM vehicle_options vo
    JOIN engines e ON vo.engine_id = e.id
    JOIN transmissions t ON vo.transmission_id = t.id
    WHERE vo.year = p_year
    AND vo.make = p_make
    AND vo.model = p_model
    AND vo.trim = p_trim
    AND t.type = p_trans_type
    ORDER BY e.name ASC;
END;
$$ LANGUAGE plpgsql;

COMMENT ON TABLE vehicle_options IS 'Denormalized table optimized for cascading dropdown queries';
COMMENT ON TABLE engines IS 'Engine specifications with detailed technical data';
COMMENT ON TABLE transmissions IS 'Transmission specifications';
COMMENT ON VIEW available_years IS 'Returns all distinct years available in the database';
COMMENT ON VIEW makes_by_year IS 'Returns makes grouped by year for dropdown population';
COMMENT ON VIEW models_by_year_make IS 'Returns models grouped by year and make';
COMMENT ON VIEW trims_by_year_make_model IS 'Returns trims grouped by year, make, and model';
COMMENT ON VIEW complete_vehicle_configs IS 'Complete vehicle configurations with all details';