180 lines
8.4 KiB
MySQL
180 lines
8.4 KiB
MySQL
-- =============================================================================
|
|
-- HOTEL RESERVATIONS — DATA MART (STAR SCHEMA)
|
|
-- Target: Oracle (university lab schema)
|
|
-- =============================================================================
|
|
|
|
-- -----------------------------------------------------------------------------
|
|
-- ETL CONTROL TABLE
|
|
-- Tracks incremental load watermarks per entity.
|
|
-- -----------------------------------------------------------------------------
|
|
|
|
CREATE TABLE ETL_WATERMARK (
|
|
entity_name VARCHAR2(50) NOT NULL,
|
|
last_key NUMBER(20,0) DEFAULT 0 NOT NULL,
|
|
last_run_ts TIMESTAMP DEFAULT SYSTIMESTAMP,
|
|
CONSTRAINT pk_etl_wm PRIMARY KEY (entity_name)
|
|
);
|
|
|
|
INSERT INTO ETL_WATERMARK (entity_name, last_key) VALUES ('FACT_ROOM_BOOKING', 0);
|
|
COMMIT;
|
|
|
|
-- -----------------------------------------------------------------------------
|
|
-- STAGING TABLES
|
|
-- NiFi loads raw MySQL data here first; SCD logic runs in pure SQL after.
|
|
-- Truncated at the start of each ETL run.
|
|
-- -----------------------------------------------------------------------------
|
|
|
|
CREATE TABLE STG_HOTEL (
|
|
hotel_id NUMBER(10,0) NOT NULL,
|
|
chain_code VARCHAR2(10),
|
|
country_code CHAR(2) NOT NULL,
|
|
star_code NUMBER(1,0) NOT NULL,
|
|
code VARCHAR2(20) NOT NULL,
|
|
name VARCHAR2(150) NOT NULL,
|
|
city VARCHAR2(100) NOT NULL
|
|
);
|
|
|
|
-- -----------------------------------------------------------------------------
|
|
-- DIMENSION TABLES
|
|
-- -----------------------------------------------------------------------------
|
|
|
|
-- YYYYMMDD integer key — cheap date range predicates, no JOIN to calendar needed
|
|
CREATE TABLE DIM_DATE (
|
|
date_key NUMBER(8,0) NOT NULL,
|
|
full_date DATE NOT NULL,
|
|
year NUMBER(4,0) NOT NULL,
|
|
quarter NUMBER(1,0) NOT NULL,
|
|
month NUMBER(2,0) NOT NULL,
|
|
month_name VARCHAR2(10) NOT NULL,
|
|
week_number NUMBER(2,0) NOT NULL,
|
|
day_of_month NUMBER(2,0) NOT NULL,
|
|
day_name VARCHAR2(10) NOT NULL,
|
|
is_weekend NUMBER(1,0) NOT NULL,
|
|
is_business_day NUMBER(1,0) NOT NULL,
|
|
season VARCHAR2(10) NOT NULL,
|
|
CONSTRAINT pk_dim_date PRIMARY KEY (date_key),
|
|
CONSTRAINT ck_dim_date_wknd CHECK (is_weekend IN (0,1)),
|
|
CONSTRAINT ck_dim_date_bday CHECK (is_business_day IN (0,1))
|
|
);
|
|
|
|
-- SCD Type 1 — country attributes are stable; just overwrite if anything changes
|
|
CREATE TABLE DIM_COUNTRY (
|
|
country_key NUMBER(10,0) GENERATED ALWAYS AS IDENTITY,
|
|
country_id NUMBER(10,0) NOT NULL,
|
|
code CHAR(2) NOT NULL,
|
|
name VARCHAR2(100) NOT NULL,
|
|
currency VARCHAR2(10) NOT NULL,
|
|
CONSTRAINT pk_dim_country PRIMARY KEY (country_key),
|
|
CONSTRAINT uq_dim_cntry_id UNIQUE (country_id)
|
|
);
|
|
|
|
-- SCD Type 1 — star rating lookup, never changes
|
|
CREATE TABLE DIM_STAR_RATING (
|
|
star_rating_key NUMBER(10,0) GENERATED ALWAYS AS IDENTITY,
|
|
star_rating_id NUMBER(10,0) NOT NULL,
|
|
code NUMBER(1,0) NOT NULL,
|
|
description VARCHAR2(20) NOT NULL,
|
|
CONSTRAINT pk_dim_star PRIMARY KEY (star_rating_key),
|
|
CONSTRAINT uq_dim_star_id UNIQUE (star_rating_id)
|
|
);
|
|
|
|
-- SCD Type 1 — chain name/code rarely changes; overwrite
|
|
CREATE TABLE DIM_HOTEL_CHAIN (
|
|
hotel_chain_key NUMBER(10,0) GENERATED ALWAYS AS IDENTITY,
|
|
hotel_chain_id NUMBER(10,0) NOT NULL,
|
|
code VARCHAR2(10) NOT NULL,
|
|
name VARCHAR2(100) NOT NULL,
|
|
CONSTRAINT pk_dim_chain PRIMARY KEY (hotel_chain_key),
|
|
CONSTRAINT uq_dim_chain_id UNIQUE (hotel_chain_id)
|
|
);
|
|
|
|
-- SCD Type 2 — hotels can change star rating or chain affiliation over time.
|
|
-- source_hotel_id is the natural key from MySQL; hotel_key is the surrogate.
|
|
-- One hotel can have multiple rows; IS_CURRENT=1 row is the active version.
|
|
-- FACT_ROOM_BOOKING links to the hotel version current at check-in date.
|
|
CREATE TABLE DIM_HOTEL (
|
|
hotel_key NUMBER(10,0) GENERATED ALWAYS AS IDENTITY,
|
|
source_hotel_id NUMBER(10,0) NOT NULL,
|
|
hotel_chain_key NUMBER(10,0),
|
|
country_key NUMBER(10,0) NOT NULL,
|
|
star_rating_key NUMBER(10,0) NOT NULL,
|
|
code VARCHAR2(20) NOT NULL,
|
|
name VARCHAR2(150) NOT NULL,
|
|
city VARCHAR2(100) NOT NULL,
|
|
-- SCD2 versioning
|
|
effective_date DATE NOT NULL,
|
|
expiry_date DATE,
|
|
is_current NUMBER(1,0) DEFAULT 1 NOT NULL,
|
|
CONSTRAINT pk_dim_hotel PRIMARY KEY (hotel_key),
|
|
CONSTRAINT ck_dh_current CHECK (is_current IN (0,1)),
|
|
CONSTRAINT fk_dh_chain FOREIGN KEY (hotel_chain_key) REFERENCES DIM_HOTEL_CHAIN (hotel_chain_key),
|
|
CONSTRAINT fk_dh_country FOREIGN KEY (country_key) REFERENCES DIM_COUNTRY (country_key),
|
|
CONSTRAINT fk_dh_star FOREIGN KEY (star_rating_key) REFERENCES DIM_STAR_RATING (star_rating_key)
|
|
);
|
|
|
|
-- SCD Type 1 — room type/floor rarely changes; upsert is sufficient
|
|
CREATE TABLE DIM_ROOM (
|
|
room_key NUMBER(10,0) GENERATED ALWAYS AS IDENTITY,
|
|
room_id NUMBER(10,0) NOT NULL,
|
|
hotel_key NUMBER(10,0) NOT NULL,
|
|
room_number VARCHAR2(10) NOT NULL,
|
|
floor NUMBER(3,0) NOT NULL,
|
|
room_type_code VARCHAR2(20) NOT NULL,
|
|
room_type_desc VARCHAR2(100) NOT NULL,
|
|
smoking_yn NUMBER(1,0) NOT NULL,
|
|
standard_rate NUMBER(10,2) NOT NULL,
|
|
CONSTRAINT pk_dim_room PRIMARY KEY (room_key),
|
|
CONSTRAINT uq_dim_room_id UNIQUE (room_id),
|
|
CONSTRAINT fk_dr_hotel FOREIGN KEY (hotel_key) REFERENCES DIM_HOTEL (hotel_key),
|
|
CONSTRAINT ck_dim_room_smk CHECK (smoking_yn IN (0,1))
|
|
);
|
|
|
|
-- SCD Type 1 — guest contact details are overwritten if they change
|
|
CREATE TABLE DIM_GUEST (
|
|
guest_key NUMBER(10,0) GENERATED ALWAYS AS IDENTITY,
|
|
guest_id NUMBER(10,0) NOT NULL,
|
|
country_key NUMBER(10,0),
|
|
name VARCHAR2(150) NOT NULL,
|
|
city VARCHAR2(100),
|
|
CONSTRAINT pk_dim_guest PRIMARY KEY (guest_key),
|
|
CONSTRAINT uq_dim_guest_id UNIQUE (guest_id),
|
|
CONSTRAINT fk_dg_country FOREIGN KEY (country_key) REFERENCES DIM_COUNTRY (country_key)
|
|
);
|
|
|
|
-- -----------------------------------------------------------------------------
|
|
-- FACT TABLE
|
|
-- -----------------------------------------------------------------------------
|
|
|
|
-- Grain: one row per room_booking.
|
|
-- source_rb_id: natural key from MySQL — used for idempotent incremental loads.
|
|
-- hotel_key: points to the DIM_HOTEL version active at check-in (SCD2 lookup).
|
|
CREATE TABLE FACT_ROOM_BOOKING (
|
|
fact_id NUMBER(10,0) GENERATED ALWAYS AS IDENTITY,
|
|
source_rb_id NUMBER(10,0) NOT NULL,
|
|
-- dimension FKs
|
|
hotel_key NUMBER(10,0) NOT NULL,
|
|
hotel_chain_key NUMBER(10,0),
|
|
room_key NUMBER(10,0) NOT NULL,
|
|
guest_key NUMBER(10,0) NOT NULL,
|
|
country_key NUMBER(10,0),
|
|
star_rating_key NUMBER(10,0) NOT NULL,
|
|
checkin_date_key NUMBER(8,0) NOT NULL,
|
|
checkout_date_key NUMBER(8,0) NOT NULL,
|
|
-- degenerate dimension
|
|
booking_status VARCHAR2(20) NOT NULL,
|
|
-- measures
|
|
nights_stayed NUMBER(4,0) NOT NULL,
|
|
nightly_rate NUMBER(10,2) NOT NULL,
|
|
total_amount NUMBER(12,2) NOT NULL,
|
|
CONSTRAINT pk_fact_rb PRIMARY KEY (fact_id),
|
|
CONSTRAINT uq_fact_rb_src UNIQUE (source_rb_id),
|
|
CONSTRAINT fk_frb_hotel FOREIGN KEY (hotel_key) REFERENCES DIM_HOTEL (hotel_key),
|
|
CONSTRAINT fk_frb_chain FOREIGN KEY (hotel_chain_key) REFERENCES DIM_HOTEL_CHAIN (hotel_chain_key),
|
|
CONSTRAINT fk_frb_room FOREIGN KEY (room_key) REFERENCES DIM_ROOM (room_key),
|
|
CONSTRAINT fk_frb_guest FOREIGN KEY (guest_key) REFERENCES DIM_GUEST (guest_key),
|
|
CONSTRAINT fk_frb_country FOREIGN KEY (country_key) REFERENCES DIM_COUNTRY (country_key),
|
|
CONSTRAINT fk_frb_star FOREIGN KEY (star_rating_key) REFERENCES DIM_STAR_RATING (star_rating_key),
|
|
CONSTRAINT fk_frb_checkin FOREIGN KEY (checkin_date_key) REFERENCES DIM_DATE (date_key),
|
|
CONSTRAINT fk_frb_checkout FOREIGN KEY (checkout_date_key) REFERENCES DIM_DATE (date_key)
|
|
);
|