def transform_data_func(df):
try:
# Removes duplicate rows based on all columns.
df = df.drop_duplicates()
# Load order_id from orders
connection = storage.engine_connect()
filter_df = pd.DataFrame(
connection.execute(
'SELECT DISTINCT order_id FROM olist_db.olist_orders_dataset;')
)
# Filter by zip_code_prefix
df = df[df.order_id.isin(filter_df[0])]
# Load product_id from sellers
connection = storage.engine_connect()
filter_df = pd.DataFrame(
connection.execute(
'SELECT DISTINCT product_id FROM olist_db.olist_products_dataset;'
))
# Filter by zip_code_prefix
df = df[df.product_id.isin(filter_df[0])]
# Load seller_id from sellers
connection = storage.engine_connect()
filter_df = pd.DataFrame(
connection.execute(
'SELECT DISTINCT seller_id FROM olist_db.olist_sellers_dataset;'
))
# Filter by zip_code_prefix
df = df[df.seller_id.isin(filter_df[0])]
return df
except BaseException as e:
raise ValueError(e)
def query_execute(query, table_name):
try:
with storage.engine_connect().begin() as connection:
connection.execute(query)
result = connection.execute('SELECT COUNT(*) FROM olist_db.{db_table_name};'.format(db_table_name = table_name))
print('Row count: ' + str([{value for value in row} for row in result if result is not None][0]))
except BaseException as e:
raise ValueError(e)
def transform_data_func(df):
try:
# Remove duplicated primary key
df = df.drop_duplicates(subset=['seller_id'])
# Load zip_code_prefix from geolocation
connection = storage.engine_connect()
filter_df = pd.DataFrame(connection.execute('SELECT DISTINCT geolocation_zip_code_prefix FROM olist_db.olist_geolocation_dataset;'))
# Filter by zip_code_prefix
df = df[df.seller_zip_code_prefix.isin(filter_df[0])]
return df
except BaseException as e:
raise ValueError(e)
def transform_data_func(df):
try:
# Remove duplicated primary key
df = df.drop_duplicates(subset=['order_id'])
# Load customer_id from customers
connection = storage.engine_connect()
filter_df = pd.DataFrame(
connection.execute(
'SELECT DISTINCT customer_id FROM olist_db.olist_customers_dataset;'
))
# Filter by zip_code_prefix
df = df[df.customer_id.isin(filter_df[0])]
return df
except BaseException as e:
raise ValueError(e)
def sales_fact(ds, **kwargs):
# CREATE TEMPORARY TABLES
query = '''
DROP TABLE IF EXISTS olist_db.temp_city;
CREATE TABLE olist_db.temp_city
SELECT
location.city_id,
location.state_id,
location.city,
location.state,
customers_dataset.customer_id
FROM
(SELECT city_id, location_state.state_id, city, state FROM olist_db.d_city AS location_city
INNER JOIN olist_db.d_state AS location_state ON location_city.state_id = location_state.state_id) AS location,
(SELECT customers_dataset.customer_id, customer_city AS city, customer_state AS state FROM olist_db.olist_customers_dataset AS customers_dataset
INNER JOIN olist_db.olist_orders_dataset AS orders_datase ON orders_datase.customer_id = customers_dataset.customer_id) AS customers_dataset
WHERE customers_dataset.city = location.city
AND customers_dataset.state = location.state;
'''
print('temp_city')
query_execute(query,'temp_city')
query = '''
DROP TABLE IF EXISTS olist_db.temp_payment;
SET @rownr=0;
CREATE TABLE olist_db.temp_payment
SELECT @rownr:=@rownr+1 AS payment_id, type_id, order_id, payment_sequential, payment_installments, payment_value FROM olist_db.olist_order_payments_dataset AS payments_dataset
INNER JOIN olist_db.d_payment_type AS payment_type ON payment_type.payment_type = payments_dataset.payment_type;
'''
print('temp_payment')
query_execute(query,'temp_payment')
# LOAD FACT SALES
f_sales = pd.DataFrame()
with storage.engine_connect().begin() as connection:
f_sales = pd.read_sql("""
SELECT
orders_dataset.order_id
, product_id
, city_id
, payment_id
, review_id
, (SELECT hour_id FROM olist_db.d_hour WHERE hour = HOUR(order_approved_at)) AS hour_id
, (SELECT day_id FROM olist_db.d_day WHERE day = DAY(order_approved_at)) AS day_id
, (SELECT month_id FROM olist_db.d_month WHERE month = MONTH(order_approved_at)) AS month_id
, (SELECT year_id FROM olist_db.d_year WHERE year = YEAR(order_approved_at)) AS year_id
, order_items_dataset.price
FROM
olist_db.olist_orders_dataset AS orders_dataset
INNER JOIN olist_db.olist_order_items_dataset AS order_items_dataset ON order_items_dataset.order_id = orders_dataset.order_id
INNER JOIN olist_db.temp_payment AS temp_payment ON temp_payment.order_id = orders_dataset.order_id
INNER JOIN olist_db.olist_customers_dataset AS customers_dataset ON customers_dataset.customer_id = orders_dataset.customer_id
INNER JOIN olist_db.temp_city AS temp_city ON temp_city.customer_id = customers_dataset.customer_id
INNER JOIN olist_db.olist_order_reviews_dataset AS olist_order_reviews_dataset ON olist_order_reviews_dataset.order_id = orders_dataset.order_id
WHERE order_approved_at IS NOT NULL;
""", connection)
storage.load_data_into_db(f_sales, 'olist_db', 'f_sales')
# DROP TEMP TABLES
query = '''
DROP TABLE IF EXISTS olist_db.temp_city;
DROP TABLE IF EXISTS olist_db.temp_payment;
'''
query_execute(query,'f_sales')
